database/openGauss-server
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
29e808dcd110d39bee7f2f1ffdacb3fdbd6c67a5
openGauss-server/src/gausskernel/optimizer/path/costsize.cpp
TotaJ 29e808dcd1 Fix parallel join cost size bug.
2020-12-09 09:04:07 +08:00

6374 lines
256 KiB
C++
Raw Blame History

/* -------------------------------------------------------------------------
*
* costsize.cpp
* Routines to compute (and set) relation sizes and path costs
*
* Path costs are measured in arbitrary units established by these basic
* parameters:
*
* seq_page_cost Cost of a sequential page fetch
* random_page_cost Cost of a non-sequential page fetch
* cpu_tuple_cost Cost of typical CPU time to process a tuple
* cpu_index_tuple_cost Cost of typical CPU time to process an index tuple
* cpu_operator_cost Cost of CPU time to execute an operator or function
* parallel_tuple_cost Cost of CPU time to pass a tuple from worker to master backend
* parallel_setup_cost Cost of setting up shared memory for parallelism
*
* We expect that the kernel will typically do some amount of read-ahead
* optimization; this in conjunction with seek costs means that seq_page_cost
* is normally considerably less than random_page_cost. (However, if the
* database is fully cached in RAM, it is reasonable to set them equal.)
*
* We also use a rough estimate "g_instance.cost_cxt.effective_cache_size" of the number of
* disk pages in Postgres + OS-level disk cache. (We can't simply use
* NBuffers for this purpose because that would ignore the effects of
* the kernel's disk cache.)
*
* Obviously, taking constants for these values is an oversimplification,
* but it's tough enough to get any useful estimates even at this level of
* detail. Note that all of these parameters are user-settable, in case
* the default values are drastically off for a particular platform.
*
* seq_page_cost and random_page_cost can also be overridden for an individual
* tablespace, in case some data is on a fast disk and other data is on a slow
* disk. Per-tablespace overrides never apply to temporary work files such as
* an external sort or a materialize node that overflows work_mem.
*
* We compute two separate costs for each path:
* total_cost: total estimated cost to fetch all tuples
* startup_cost: cost that is expended before first tuple is fetched
* In some scenarios, such as when there is a LIMIT or we are implementing
* an EXISTS(...) sub-select, it is not necessary to fetch all tuples of the
* path's result. A caller can estimate the cost of fetching a partial
* result by interpolating between startup_cost and total_cost. In detail:
* actual_cost = startup_cost +
* (total_cost - startup_cost) * tuples_to_fetch / path->rows;
* Note that a base relation's rows count (and, by extension, plan_rows for
* plan nodes below the LIMIT node) are set without regard to any LIMIT, so
* that this equation works properly. (Also, these routines guarantee not to
* set the rows count to zero, so there will be no zero divide.) The LIMIT is
* applied as a top-level plan node.
*
* For largely historical reasons, most of the routines in this module use
* the passed result Path only to store their results (rows, startup_cost and
* total_cost) into. All the input data they need is passed as separate
* parameters, even though much of it could be extracted from the Path.
* An exception is made for the cost_XXXjoin() routines, which expect all
* the other fields of the passed XXXPath to be filled in, and similarly
* cost_index() assumes the passed IndexPath is valid except for its output
* values.
*
* Portions Copyright (c) 2020 Huawei Technologies Co.,Ltd.
* Portions Copyright (c) 1996-2012, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* src/gausskernel/optimizer/path/costsize.cpp
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "knl/knl_variable.h"
#include <math.h>
#include "catalog/pg_partition_fn.h"
#include "catalog/pg_proc.h"
#include "executor/executor.h"
#include "executor/hashjoin.h"
#include "executor/nodeHash.h"
#include "miscadmin.h"
#include "nodes/nodeFuncs.h"
#include "optimizer/bucketpruning.h"
#include "optimizer/clauses.h"
#include "optimizer/cost.h"
#include "optimizer/optimizerdebug.h"
#include "optimizer/pathnode.h"
#include "optimizer/paths.h"
#include "optimizer/placeholder.h"
#include "optimizer/plancat.h"
#include "optimizer/planmain.h"
#include "optimizer/restrictinfo.h"
#include "optimizer/tlist.h"
#include "parser/parsetree.h"
#include "utils/dynahash.h"
#include "utils/guc.h"
#include "utils/lsyscache.h"
#include "utils/selfuncs.h"
#include "utils/spccache.h"
#include "utils/tuplesort.h"
#include "catalog/pg_aggregate.h"
#include "catalog/pg_operator.h"
#include "catalog/pg_proc.h"
#include "vectorsonic/vsonichash.h"
#include "pgxc/pgxc.h"
/*
* For columnar scan, the cpu cost is less than seq scan, so
* we are definition that the cost of scanning one tuple is 1/10 times.
*/
#define COL_TUPLE_COST_MULTIPLIER 10
/*
* Estimate the overhead per hashtable entry at 64 bytes (same as in
* planner.c).
*/
#define HASH_ENTRY_OVERHEAD 64
/* The default value of the column row width */
#define COL_TUPLE_WIDTH 30
typedef struct {
PlannerInfo* root;
QualCost total;
} cost_qual_eval_context;
/* Identify the max global rows of joinrel if have over estimate. */
#define JOINREL_MAX_GLOBAL_ROWS (double)(1.0e11)
static bool cost_qual_eval_walker(Node* node, cost_qual_eval_context* context);
static void get_restriction_qual_cost(
PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info, QualCost* qpqual_cost);
static double calc_joinrel_size_estimate(PlannerInfo* root, double outer_rows, double inner_rows,
SpecialJoinInfo* sjinfo, List* restrictlist, bool varratio_cached);
static int calc_distributekey_width(Path* path, int* width, bool vectorized, bool aligned);
static Cost get_subqueryscan_stream_cost(Plan* subplan);
static double get_parallel_divisor(Path *path);
extern int getDataMinLen(Oid typeOid, int typeMod);
extern bool isExprSonicEnable(Expr* node);
extern bool isAggrefSonicEnable(Oid aggfnoid);
static Cost append_nonpartial_cost(List *subpaths, int numpaths, int parallel_workers);
/*
* init_plan_cost
* used to initialize a plan's cost related field
*
* @param (in) plan:
* the plan to be handled
*
* @return: void
*/
void init_plan_cost(Plan* plan)
{
plan->startup_cost = 0.0;
plan->total_cost = 0.0;
plan->multiple = 1.0;
plan->plan_rows = 0.0;
plan->plan_width = 0;
plan->innerdistinct = 1.0;
plan->outerdistinct = 1.0;
plan->pred_rows = -1.0;
plan->pred_startup_time = -1.0;
plan->pred_total_time = -1.0;
plan->pred_max_memory = -1;
plan->parallel_aware = false;
plan->parallel_safe = false;
}
static inline void get_info_from_rel(
Relation relation, int* maxBatchRow, bool* isPartTable, bool* isValuePartTable, int* partialClusterRows)
{
*maxBatchRow = RelationGetMaxBatchRows(relation);
*isPartTable = RELATION_IS_PARTITIONED(relation);
*isValuePartTable = RELATION_IS_VALUE_PARTITIONED(relation);
*partialClusterRows = RelationGetPartialClusterRows(relation);
}
/*
* Description: estimation the memory info for cstoreinsert and dfsinsert .
*
* For dfs insert : maxMem = insert memory + sort memory(if has pck). If the table is value partition,
* we set 2G memory for all dynamic partitions to use. If the table is not partition table,we shoule used
* 128MB to promise write disk.
*
* For cstore insert : maxMem = insert memory + sort memory(if has pck). insetMem = maxBatchrow * column *3.
* the means of 3 is that the insert will need three part memory for batchinsert\beforecompress\aftercompress.
* sortMem = partialClusterRows(tuples) * column. Default partialClusterRows is 420w. if tuples number is less than
* 420w, the true values(tuples number) is used to caculate the memory. 1) pck+partition: insetMem is set to 2g and sort
* memory is estimated to 4g. 2) partition: insetMem is maxBatchrow * column *3. insetMem<=2g. Default maxBatchrow is 6w.
* 3) pck+ table: insertMem is maxBatchrow * column*3. sortMem is partialClusterRows(tuples) * column.
* 4) table: insertMem is the total memory, is maxBatchrow * column*3
*
* Parameters:
* @in path: the path for insert.
* @in root: plannerinfo struct for current query level.
* @in input_cost: is the total cost for reading the input data.
* @in tuples: is the number of tuples in the relation.
* @in width: is the average tuple width in bytes.
* @in comparison_cost: is the extra cost per comparison, if any.
* @in modify_mem: is the number of kilobytes of work memory allowed for the sort.
* @in dop: set the dop.
* @in resultRelOid: the relation to be scanned.
* @in isDfsStore: judge whether table is dfs table.
* @in mem_info: is operator max and min info used by memory control module.
* Return: void
*/
void cost_insert(Path* path, bool vectorized, Cost input_cost, double tuples, int width, Cost comparison_cost,
int modify_mem, int dop, Oid resultRelOid, bool isDfsStore, OpMemInfo* mem_info)
{
Cost startup_cost = input_cost;
Cost run_cost = 0;
double input_bytes = relation_byte_size(tuples, width, vectorized) / SET_DOP(dop);
double output_bytes = 0;
double output_bytes_insert = 0;
double output_bytes_pck = 0;
long modify_mem_bytes = modify_mem * 1024L / SET_DOP(dop);
int partitionNum = 1;
/* isPartTable is judge whether is range partition. isValuePartTable is judge whether is value partition */
bool isPartTable = false;
bool isValuePartTable = false;
bool hasPck = false;
Relation relation;
int maxBatchRow = MAX_BATCH_ROWS;
int partialClusterRows = PARTIAL_CLUSTER_ROWS;
double sortRows = 1;
/* We should compute the table's partition num and maxBatchRow and pck and index information. */
if (resultRelOid) {
relation = relation_open(resultRelOid, AccessShareLock);
get_info_from_rel(relation, &maxBatchRow, &isPartTable, &isValuePartTable, &partialClusterRows);
if (isPartTable) {
partitionNum = getNumberOfRangePartitions(relation);
}
if (relation->rd_rel->relhasclusterkey) {
hasPck = true;
}
relation_close(relation, NoLock);
}
/*
* We want to be sure the cost of a sort is never estimated as zero, even
* if passed-in tuple count is zero. Besides, mustn't do log(0)...
*/
if (tuples < 2.0) {
tuples = 2.0;
}
/* Include the default cost-per-comparison */
comparison_cost += 2.0 * u_sess->attr.attr_sql.cpu_operator_cost;
/* if dfs table for insert, the memory is 128MB. If cstore table for insert, the memory is maxBatchRow*width. */
if (isDfsStore)
output_bytes_insert = DFS_MIN_MEM_SIZE * MEM_KB;
else
output_bytes_insert = relation_byte_size(maxBatchRow, width, vectorized) * 3;
if (output_bytes_insert > modify_mem_bytes) {
/* CPU costs : Assume about N log2 N comparisons */
startup_cost += comparison_cost * tuples * LOG2(tuples);
/* Disk costs */
startup_cost += compute_sort_disk_cost(input_bytes, modify_mem_bytes);
} else {
if (tuples > 2 * maxBatchRow || input_bytes > modify_mem_bytes) {
/*
* We'll use a bounded heap-sort keeping just K tuples in memory, for
* a total number of tuple comparisons of N log2 K; but the constant
* factor is a bit higher than for quicksort. Tweak it so that the
* cost curve is continuous at the crossover point.
*/
startup_cost += comparison_cost * tuples * LOG2(2.0 * maxBatchRow);
} else {
/* We'll use plain quicksort on all the input tuples */
startup_cost += comparison_cost * tuples * LOG2(tuples);
}
}
if (isDfsStore) {
sortRows = tuples > partialClusterRows ? partialClusterRows : tuples;
if (mem_info != NULL) {
mem_info->opMem = modify_mem;
if (hasPck && isValuePartTable) {
output_bytes_insert = PARTITION_MAX_SIZE * MEM_KB;
output_bytes_pck = relation_byte_size(sortRows, width, vectorized);
mem_info->maxMem = output_bytes_pck / MEM_KB + output_bytes_insert / MEM_KB;
mem_info->minMem = mem_info->maxMem;
} else if (!hasPck && isValuePartTable) {
output_bytes = PARTITION_MAX_SIZE * MEM_KB;
mem_info->maxMem = output_bytes / MEM_KB;
mem_info->minMem = mem_info->maxMem;
} else if (hasPck && !isValuePartTable) {
output_bytes_pck = relation_byte_size(sortRows, width, vectorized);
mem_info->maxMem = (output_bytes_insert + output_bytes_pck) / MEM_KB;
mem_info->minMem = output_bytes_insert / MEM_KB + output_bytes_pck / MEM_KB / SORT_MAX_DISK_SIZE;
} else {
mem_info->maxMem = output_bytes_insert / MEM_KB;
mem_info->minMem = output_bytes_insert / MEM_KB;
}
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
MEMCTL_LOG(DEBUG2,
"DFS INSERT:The opMem is: %lfKB, the maxMem is: %lfKB, the minMem is: %lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
}
} else {
/* calucate the mem_info for partition\ partition_pck,cstoretable\cstoretable_pck.*/
if (mem_info != NULL) {
mem_info->opMem = modify_mem;
if (hasPck && isPartTable) {
/* we will need 2g memory to insert ,4g memory to sort. NOTICE : PARTITION_MAX_SIZE is KB */
output_bytes_pck = PARTITION_MAX_SIZE * MEM_KB * 2;
output_bytes_insert = PARTITION_MAX_SIZE * MEM_KB;
mem_info->maxMem = (output_bytes_pck + output_bytes_insert) / MEM_KB;
} else if (!hasPck && isPartTable) {
output_bytes = output_bytes_insert;
double output_k_bytes = output_bytes / MEM_KB;
mem_info->maxMem = (output_k_bytes > PARTITION_MAX_SIZE) ? output_k_bytes : PARTITION_MAX_SIZE;
} else if (hasPck && !isPartTable) {
output_bytes_pck = relation_byte_size(partialClusterRows, width, vectorized);
output_bytes = output_bytes_pck + output_bytes_insert;
mem_info->maxMem = output_bytes / MEM_KB;
} else {
output_bytes = output_bytes_insert;
mem_info->maxMem = output_bytes / MEM_KB;
}
mem_info->minMem = mem_info->maxMem / SORT_MAX_DISK_SIZE;
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
MEMCTL_LOG(DEBUG2,
"CSTORE INSERT:The opMem is: %lfKB, the maxMem is: %lfKB, the minMem is: %lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
}
}
/*
* Also charge a small amount (arbitrarily set equal to operator cost) per
* extracted tuple. We don't charge cpu_tuple_cost because a Sort node
* doesn't do qual-checking or projection, so it has less overhead than
* most plan nodes. Note it's correct to use tuples not output_tuples
* here --- the upper LIMIT will pro-rate the run cost so we'd be double
* counting the LIMIT otherwise.
*/
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * tuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
}
/*
* Description: estimation the memory info for cstoredelete and dfsdelete .
*
* For delete : maxMem = delete mem(for sort). It is same between cstoredelete and dfsdelete .
* sortMem = partialClusterRows(tuples) * column. If the mem is less than 16MB,maxMem >=16MB.
* Default partialClusterRows is 420w. if tuples number is less than 420w, the true values(tuples number)
* is used to caculate the memory.
*
* Parameters:
* @in path: the path for delete.
* @in root: plannerinfo struct for current query level.
* @in input_cost: is the total cost for reading the input data.
* @in tuples: is the number of tuples in the relation.
* @in width: is the average tuple width in bytes.
* @in comparison_cost: is the extra cost per comparison, if any.
* @in modify_mem: is the number of kilobytes of work memory allowed for the sort.
* @in dop: set the dop.
* @in resultRelOid: the relation to be scanned.
* @in isDfsStore: judge whether table is dfs table.
* @in mem_info: is operator max and min info used by memory control module.
* Return: void
*/
void cost_delete(Path* path, bool vectorized, Cost input_cost, double tuples, int width, Cost comparison_cost,
int modify_mem, int dop, Oid resultRelOid, bool isDfsStore, OpMemInfo* mem_info)
{
Cost startup_cost = input_cost;
Cost run_cost = 0;
double input_bytes = relation_byte_size(tuples, width, vectorized) / SET_DOP(dop);
double output_bytes = 0;
double output_tuples = 0;
long modify_mem_bytes = modify_mem * 1024L / SET_DOP(dop);
Relation relation;
int partialClusterRows = PARTIAL_CLUSTER_ROWS;
if (resultRelOid) {
relation = relation_open(resultRelOid, AccessShareLock);
partialClusterRows = RelationGetPartialClusterRows(relation);
relation_close(relation, NoLock);
}
/*
* We want to be sure the cost of a sort is never estimated as zero, even
* if passed-in tuple count is zero. Besides, mustn't do log(0)...
*/
if (tuples < 2.0) {
tuples = 2.0;
}
/* Include the default cost-per-comparison */
comparison_cost += 2.0 * u_sess->attr.attr_sql.cpu_operator_cost;
output_tuples = tuples > partialClusterRows ? partialClusterRows : tuples;
output_bytes = relation_byte_size(output_tuples, width, vectorized);
if (output_bytes > modify_mem_bytes) {
/* CPU costs : Assume about N log2 N comparisons */
startup_cost += comparison_cost * tuples * LOG2(tuples);
/* Disk costs */
startup_cost += compute_sort_disk_cost(input_bytes, modify_mem_bytes);
} else {
if (tuples > 2 * output_tuples || input_bytes > modify_mem_bytes) {
/*
* We'll use a bounded heap-sort keeping just K tuples in memory, for
* a total number of tuple comparisons of N log2 K; but the constant
* factor is a bit higher than for quicksort. Tweak it so that the
* cost curve is continuous at the crossover point.
*/
startup_cost += comparison_cost * tuples * LOG2(2.0 * output_tuples);
} else {
/* We'll use plain quicksort on all the input tuples */
startup_cost += comparison_cost * tuples * LOG2(tuples);
}
}
/*
* calucate the mem_info for cstore table or dfs table.
*/
if (mem_info != NULL) {
mem_info->opMem = modify_mem;
mem_info->maxMem = output_bytes / MEM_KB > SORT_MIM_MEM ? output_bytes / MEM_KB : SORT_MIM_MEM;
mem_info->minMem = mem_info->maxMem / SORT_MAX_DISK_SIZE;
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
MEMCTL_LOG(DEBUG2,
"MEMORY DELETE:The opMem is : %lfKB, the maxMem is :%lfKB, the minMem is :%lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
}
/*
* Also charge a small amount (arbitrarily set equal to operator cost) per
* extracted tuple. We don't charge cpu_tuple_cost because a Sort node
* doesn't do qual-checking or projection, so it has less overhead than
* most plan nodes. Note it's correct to use tuples not output_tuples
* here --- the upper LIMIT will pro-rate the run cost so we'd be double
* counting the LIMIT otherwise.
*/
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * tuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
}
/*
* Description: estimation the memory info for cstoreupdate and dfsupdate . update = delete +insert(insert + sort).
* So, we should caculate the delete mem(sort mem) and insert mem(insert and sort). Here, the deleted
* memory can be reused for the pck sort memory when insert, so regardless of whether there is a pck,
* we need to calculate the memory required by sort. So, update mem = sortMem(delete) + insertMem.
*
* For dfs update : maxMem = insert memory + sort memory. For insertMem: If the table is value partition,
* we set 2G memory for all dynamic partitions to use. If the table is not partition table,we shoule used
* 128MB to promise write disk. For sortMem: partialClusterRows(tuples) * column. It is important to
* note that the minimum value needs to >= 128MB.
*
* For cstore update : maxMem = insert memory + sort memory. insetMem = maxBatchrow * column *3.
* the means of 3 is that the insert will need three part memory for batchinsert\beforecompress\aftercompress.
* sortMem = partialClusterRows(tuples) * column. Default partialClusterRows is 420w. if tuples number is less than
* 420w, the true values(tuples number) is used to caculate the memory. 1) partition: insetMem is set to 2g and sort
* memory is estimated to 4g. 2) table: insertMem is maxBatchrow * column*3. sortMem is partialClusterRows(tuples) *
* column.
*
* Parameters:
* @in path: the path for update.
* @in root: plannerinfo struct for current query level.
* @in input_cost: is the total cost for reading the input data.
* @in tuples: is the number of tuples in the relation.
* @in width: is the average tuple width in bytes.
* @in comparison_cost: is the extra cost per comparison, if any.
* @in modify_mem: is the number of kilobytes of work memory allowed for the sort.
* @in dop: set the dop.
* @in resultRelOid: the relation to be scanned.
* @in isDfsStore: judge whether table is dfs table.
* @in mem_info: is operator max and min info used by memory control module.
* Return: void
*/
void cost_update(Path* path, bool vectorized, Cost input_cost, double tuples, int width, Cost comparison_cost,
int modify_mem, int dop, Oid resultRelOid, bool isDfsStore, OpMemInfo* mem_info)
{
Cost startup_cost = input_cost;
Cost run_cost = 0;
double input_bytes = relation_byte_size(tuples, width, vectorized) / SET_DOP(dop);
double output_bytes = 0;
double output_bytes_insert = 0;
double output_bytes_pck = 0;
long modify_mem_bytes = modify_mem * 1024L / SET_DOP(dop);
int partitionNum = 0;
bool isPartTable = false;
bool isValuePartTable = false;
bool hasPck = false;
Relation relation;
int maxBatchRow = MAX_BATCH_ROWS;
int partialClusterRows = PARTIAL_CLUSTER_ROWS;
/* We should compute the table's partition num and maxBatchRow and pck and index information. */
if (resultRelOid) {
relation = relation_open(resultRelOid, AccessShareLock);
get_info_from_rel(relation, &maxBatchRow, &isPartTable, &isValuePartTable, &partialClusterRows);
if (isPartTable) {
partitionNum = getNumberOfRangePartitions(relation);
}
if (relation->rd_rel->relhasclusterkey) {
hasPck = true;
}
relation_close(relation, NoLock);
}
/*
* We want to be sure the cost of a sort is never estimated as zero, even
* if passed-in tuple count is zero. Besides, mustn't do log(0)...
*/
if (tuples < 2.0) {
tuples = 2.0;
}
/* Include the default cost-per-comparison */
comparison_cost += 2.0 * u_sess->attr.attr_sql.cpu_operator_cost;
/* If dfs table for insert, the memory is 128MB.
* If dfs part table for insert, the memory is 2GB.
* If cstore table for insert, the memory is maxBatchRow*width.
*/
if (isDfsStore)
output_bytes_insert = isValuePartTable ? PARTITION_MAX_SIZE * MEM_KB : DFS_MIN_MEM_SIZE * MEM_KB;
else
output_bytes_insert = relation_byte_size(maxBatchRow, width, vectorized) * 3;
if (output_bytes_insert > modify_mem_bytes) {
/* CPU costs : Assume about N log2 N comparisons */
startup_cost += comparison_cost * tuples * LOG2(tuples);
/* Disk costs */
startup_cost += compute_sort_disk_cost(input_bytes, modify_mem_bytes);
} else {
if (tuples > 2 * maxBatchRow || input_bytes > modify_mem_bytes) {
/*
* We'll use a bounded heap-sort keeping just K tuples in memory, for
* a total number of tuple comparisons of N log2 K; but the constant
* factor is a bit higher than for quicksort. Tweak it so that the
* cost curve is continuous at the crossover point.
*/
startup_cost += comparison_cost * tuples * LOG2(2.0 * maxBatchRow);
} else {
/* We'll use plain quicksort on all the input tuples */
startup_cost += comparison_cost * tuples * LOG2(tuples);
}
}
/*
* calucate the mem_info for partition\ cstoretable. delete memory + insert memory.
* delete sort will be reused to insert sort.
*/
if (isDfsStore) {
if (mem_info != NULL) {
mem_info->opMem = modify_mem;
output_bytes_pck = relation_byte_size(partialClusterRows, width, vectorized);
output_bytes = output_bytes_pck + output_bytes_insert;
mem_info->maxMem = output_bytes / MEM_KB;
mem_info->minMem = output_bytes_insert / MEM_KB + output_bytes_pck / SORT_MAX_DISK_SIZE / MEM_KB;
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
MEMCTL_LOG(DEBUG2,
"DFS UPDATE:The opMem is: %lfKB, the maxMem is: %lfKB, the minMem is: %lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
}
} else {
if (mem_info != NULL) {
mem_info->opMem = modify_mem;
if (isPartTable) {
/* We will need 2g memory to insert ,4g memory to sort.*/
output_bytes_pck = PARTITION_MAX_SIZE * MEM_KB * 2;
output_bytes_insert = PARTITION_MAX_SIZE * MEM_KB;
mem_info->maxMem = (output_bytes_pck + output_bytes_insert) / MEM_KB;
mem_info->minMem = output_bytes_insert / MEM_KB + output_bytes_pck / SORT_MAX_DISK_SIZE / MEM_KB;
MEMCTL_LOG(DEBUG2,
"CSTORE PART TABLE UPDATE:The opMem is: %lfKB, the maxMem is: %lfKB, the minMem is: %lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
} else {
output_bytes_pck = relation_byte_size(partialClusterRows, width, vectorized);
output_bytes = output_bytes_pck + output_bytes_insert;
mem_info->maxMem = output_bytes / MEM_KB;
mem_info->minMem = output_bytes_insert / MEM_KB + output_bytes_pck / SORT_MAX_DISK_SIZE / MEM_KB;
MEMCTL_LOG(DEBUG2,
"CSTORE TABLE UPDATE:The opMem is: %lfKB, the maxMem is: %lfKB, the minMem is: %lfKB",
mem_info->opMem,
mem_info->maxMem,
mem_info->minMem);
}
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
}
}
/*
* Also charge a small amount (arbitrarily set equal to operator cost) per
* extracted tuple. We don't charge cpu_tuple_cost because a Sort node
* doesn't do qual-checking or projection, so it has less overhead than
* most plan nodes. Note it's correct to use tuples not output_tuples
* here --- the upper LIMIT will pro-rate the run cost so we'd be double
* counting the LIMIT otherwise.
*/
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * tuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
}
/*
* clamp_row_est
* Force a row-count estimate to a sane value.
*/
double clamp_row_est(double nrows)
{
/*
* Force estimate to be at least one row, to make explain output look
* better and to avoid possible divide-by-zero when interpolating costs.
* Make it an integer, too.
*/
if (nrows <= 1.0) {
nrows = 1.0;
} else {
nrows = rint(nrows);
}
return nrows;
}
/* Set local and global rows for sort/agg/group by/material/windowfun path for lower path. */
void set_path_rows(Path* path, double rows, double multiple)
{
path->rows = rows;
path->multiple = multiple;
}
/* Set local and global rows for baserel or joinrel path. */
void set_rel_path_rows(Path* path, RelOptInfo* rel, ParamPathInfo* param_info)
{
/* Mark the path with the correct row estimate */
if (param_info != NULL)
set_path_rows(path, param_info->ppi_rows);
else
set_path_rows(path, rel->rows, rel->multiple);
}
/*
* Set the real path rows after parallel.
*
* @in_param path: the path need to be corrected.
*/
static void set_parallel_path_rows(Path* path)
{
int dop = SET_DOP(path->dop);
/*
* When we parallel replicate path, the return rows in one node
* and global will both increase. This can be useful when we parallel
* join inner path.
*/
if (is_replicated_path(path)) {
path->rows *= dop;
}
}
/*
* cost_seqscan
* Determines and returns the cost of scanning a relation sequentially.
* The pruning ration for Partitioned table will be considered in set_plain_rel_size().
* 'baserel' is the relation to be scanned
* 'param_info' is the ParamPathInfo if this is a parameterized path, else NULL
*/
void cost_seqscan(Path* path, PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info)
{
Cost startup_cost = 0;
Cost run_cost = 0;
double spc_seq_page_cost;
QualCost qpqual_cost;
Cost cpu_per_tuple = 0.0;
int dop = SET_DOP(path->dop);
/* Should only be applied to base relations */
Assert(baserel->relid > 0);
Assert(baserel->rtekind == RTE_RELATION);
/* Mark the path with the correct row estimate */
set_rel_path_rows(path, baserel, param_info);
set_parallel_path_rows(path);
/* fetch estimated page cost for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, NULL, &spc_seq_page_cost);
get_restriction_qual_cost(root, baserel, param_info, &qpqual_cost);
startup_cost += qpqual_cost.startup;
if (!u_sess->attr.attr_sql.enable_seqscan)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* When we parallel the scan node, then the disk costs and cpu costs
* wiil be equal division to all parallelism thread.
*/
run_cost += u_sess->opt_cxt.smp_thread_cost * (dop - 1);
run_cost += spc_seq_page_cost * baserel->pages / dop;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qpqual_cost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples) / dop;
/*
* Primitive parallel cost model. Assume the leader will do half as much
* work as a regular worker, because it will also need to read the tuples
* returned by the workers when they percolate up to the gather ndoe.
* This is almost certainly not exactly the right way to model this, so
* this will probably need to be changed at some point...
*/
if (path->parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(path);
run_cost = run_cost / parallel_divisor;
path->rows = clamp_row_est(path->rows / parallel_divisor);
}
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_seqscan)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* Description: Determines and returns the cost of scanning a relation using sampling.
*
* Parameters:
* @in path: seqscan or cstorescan path.
* @in root: plannerinfo struct for current query level.
* @in baserel: the relation to be scanned.
* @in param_info: the ParamPathInfo if this is a parameterized path, else NULL
*
* Return: void
*/
void cost_samplescan(Path* path, PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info)
{
Cost startup_cost = 0;
Cost run_cost = 0;
RangeTblEntry* rte = NULL;
TableSampleClause* tsc = NULL;
double spc_seq_page_cost, spc_random_page_cost, spc_page_cost;
QualCost qpqual_cost;
Cost cpu_per_tuple = 0.0;
/* Should only be applied to base relations with tablesample clauses */
AssertEreport(baserel->relid > 0,
MOD_OPT,
"The relid is invalid when determining the cost of scanning a relation using sampling.");
rte = planner_rt_fetch(baserel->relid, root);
AssertEreport(rte->rtekind == RTE_RELATION,
MOD_OPT,
"Only base relation can be supported when determining the cost of scanning a relation using sampling.");
tsc = rte->tablesample;
AssertEreport(tsc != NULL,
MOD_OPT,
"Samling method and parameters is null when determining the cost of scanning a relation using sampling.");
set_rel_path_rows(path, baserel, param_info);
/* fetch estimated page cost for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, &spc_random_page_cost, &spc_seq_page_cost);
/* if sampleType is SYSTEM_SAMPLE, assume random access, else sequential */
spc_page_cost = (tsc->sampleType == BERNOULLI_SAMPLE) ? spc_seq_page_cost : spc_random_page_cost;
/*
* disk costs (recall that baserel->pages has already been set to the
* number of pages the sampling method will visit)
*/
run_cost += spc_page_cost * baserel->pages;
/*
* CPU costs (recall that baserel->tuples has already been set to the
* number of tuples the sampling method will select). Note that we ignore
* execution cost of the TABLESAMPLE parameter expressions; they will be
* evaluated only once per scan, and in most usages they'll likely be
* simple constants anyway. We also don't charge anything for the
* calculations the sampling method might do internally.
*/
get_restriction_qual_cost(root, baserel, param_info, &qpqual_cost);
startup_cost += qpqual_cost.startup;
if (baserel->orientation == REL_COL_ORIENTED) {
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost / COL_TUPLE_COST_MULTIPLIER + qpqual_cost.per_tuple;
} else if (baserel->orientation == REL_TIMESERIES_ORIENTED) {
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("Unsupported Using Index FOR TIMESERIES.")));
} else {
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qpqual_cost.per_tuple;
}
run_cost += cpu_per_tuple * baserel->tuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
}
/*
* cost_cstorescan
* Determines and returns the cost of scanning a column store.
* The pruning ration for Partitioned table will be considered in set_plain_rel_size().
*/
void cost_cstorescan(Path* path, PlannerInfo* root, RelOptInfo* baserel)
{
double spc_seq_page_cost;
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_per_tuple = 0.0;
int dop = SET_DOP(path->dop);
/* Should only be applied to base relations */
Assert(baserel->relid > 0 && baserel->rtekind == RTE_RELATION);
set_rel_path_rows(path, baserel, NULL);
set_parallel_path_rows(path);
/* fetch estimated page cost for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, NULL, &spc_seq_page_cost);
startup_cost += baserel->baserestrictcost.startup;
if (!u_sess->attr.attr_sql.enable_seqscan)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* When we parallel the scan node, then the disk costs and cpu costs
* wiil be equal division to all parallelism thread.
*/
run_cost += u_sess->opt_cxt.smp_thread_cost * (dop - 1);
run_cost += spc_seq_page_cost * baserel->pages / dop;
cpu_per_tuple =
u_sess->attr.attr_sql.cpu_tuple_cost / COL_TUPLE_COST_MULTIPLIER + baserel->baserestrictcost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples) / dop;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_seqscan)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* Determines and returns the cost of scanning a DFS relation.
* path: The scan path.
* root: The PlannerInfo struct.
* baserel: The relation to be scanned.
*/
void cost_dfsscan(Path* path, PlannerInfo* root, RelOptInfo* baserel)
{
Cost startup_cost = 0;
Cost run_cost = 0;
double spc_seq_page_cost;
Cost cpu_per_tuple = 0.0;
int dop = SET_DOP(path->dop);
/*
* Should only be applied to base relations.
*/
AssertEreport(
baserel->relid > 0, MOD_OPT, "The relid is invalid when determining the cost of scanning a DFS relation.");
AssertEreport(baserel->rtekind == RTE_RELATION,
MOD_OPT,
"Only base relation can be supported when determining the cost of scanning a DFS relation.");
set_rel_path_rows(path, baserel, NULL);
set_parallel_path_rows(path);
/*
* Fetch estimated page cost for tablespace containing table.
*/
get_tablespace_page_costs(baserel->reltablespace, NULL, &spc_seq_page_cost);
startup_cost = baserel->baserestrictcost.startup;
/* Parallelize start up cost. */
if (!u_sess->attr.attr_sql.enable_seqscan)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* When we parallel the scan node, then the disk costs and cpu costs
* wiil be equal division to all parallelism thread.
*/
run_cost += u_sess->opt_cxt.smp_thread_cost * (dop - 1);
cpu_per_tuple =
u_sess->attr.attr_sql.cpu_tuple_cost / COL_TUPLE_COST_MULTIPLIER + baserel->baserestrictcost.per_tuple;
run_cost += (cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples)) / dop +
(spc_seq_page_cost * baserel->pages) / dop;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_seqscan)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
/*
* data redistribution for DFS table.
*/
if (true == u_sess->attr.attr_sql.enable_cluster_resize && root->query_level == 1 &&
root->parse->commandType == CMD_INSERT) {
root->dataDestRelIndex = baserel->relid;
}
}
/*
* cost_tsstorescan
* Determines and returns the cost of scanning a time series store.
*/
void cost_tsstorescan(Path *path, PlannerInfo *root, RelOptInfo *baserel)
{
double spc_seq_page_cost;
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_per_tuple = 0.0;
int dop = SET_DOP(path->dop);
/* Should only be applied to base relations */
Assert(baserel->relid > 0 && baserel->rtekind == RTE_RELATION);
set_rel_path_rows(path, baserel, NULL);
set_parallel_path_rows(path);
/* fetch estimated page cost for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, NULL, &spc_seq_page_cost);
startup_cost += baserel->baserestrictcost.startup;
if (!u_sess->attr.attr_sql.enable_seqscan) {
startup_cost += g_instance.cost_cxt.disable_cost;
}
/*
* When we parallel the scan node, then the disk costs and cpu costs
* wiil be equal division to all parallelism thread.
*/
run_cost += u_sess->opt_cxt.smp_thread_cost * (dop - 1);
run_cost += spc_seq_page_cost * baserel->pages / dop;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost / COL_TUPLE_COST_MULTIPLIER
+ baserel->baserestrictcost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples) / dop;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_seqscan) {
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
}
/*
* cost_gather
* Determines and returns the cost of gather path.
*
* 'rel' is the relation to be operated upon
* 'param_info' is the ParamPathInfo if this is a parameterized path, else NULL
*/
void cost_gather(GatherPath *path, RelOptInfo *rel, ParamPathInfo *param_info)
{
Cost startup_cost = 0;
Cost run_cost = 0;
/* Mark the path with the correct row estimate */
if (param_info)
path->path.rows = param_info->ppi_rows;
else
path->path.rows = rel->rows;
startup_cost = path->subpath->startup_cost;
run_cost = path->subpath->total_cost - path->subpath->startup_cost;
/* Parallel setup and communication cost. */
startup_cost += u_sess->attr.attr_sql.parallel_setup_cost;
run_cost += u_sess->attr.attr_sql.parallel_tuple_cost * path->path.rows;
path->path.startup_cost = startup_cost;
path->path.total_cost = (startup_cost + run_cost);
}
/*
* cost_index
* Determines and returns the cost of scanning a relation using an index.
*
* 'path' describes the indexscan under consideration, and is complete
* except for the fields to be set by this routine
* 'loop_count' is the number of repetitions of the indexscan to factor into
* estimates of caching behavior
*
* In addition to rows, startup_cost and total_cost, cost_index() sets the
* path's indextotalcost and indexselectivity fields. These values will be
* needed if the IndexPath is used in a BitmapIndexScan.
*
* NOTE: path->indexquals must contain only clauses usable as index
* restrictions. Any additional quals evaluated as qpquals may reduce the
* number of returned tuples, but they won't reduce the number of tuples
* we have to fetch from the table, so they don't reduce the scan cost.
*/
void cost_index(IndexPath* path, PlannerInfo* root, double loop_count, bool partial_path)
{
IndexOptInfo* index = path->indexinfo;
RelOptInfo* baserel = index->rel;
bool indexonly = (path->path.pathtype == T_IndexOnlyScan);
List* allclauses = NIL;
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_run_cost = 0;
Cost indexStartupCost;
Cost indexTotalCost;
Selectivity indexSelectivity;
double indexCorrelation, csquared;
double spc_seq_page_cost, spc_random_page_cost;
Cost min_IO_cost, max_IO_cost;
QualCost qpqual_cost;
Cost cpu_per_tuple = 0.0;
double tuples_fetched;
double pages_fetched;
bool ispartitionedindex = path->indexinfo->rel->isPartitionedTable;
double rand_heap_pages;
double index_pages = 0.0;
/* Should only be applied to base relations */
AssertEreport(IsA(baserel, RelOptInfo) && IsA(index, IndexOptInfo),
MOD_OPT,
"The nodeTag of baserel is not T_RelOptInfo, or the nodeTag of index is not T_IndexOptInfo"
"when determining the cost of scanning a relation using an index.");
AssertEreport(baserel->relid > 0,
MOD_OPT,
"The relid is invalid when determining the cost of scanning a relation using an index.");
AssertEreport(baserel->rtekind == RTE_RELATION,
MOD_OPT,
"Only base relation can be supported when determining the cost of scanning a relation using an index.");
set_rel_path_rows(&path->path, baserel, path->path.param_info);
/* Mark the path with the correct row estimate */
if (path->path.param_info) {
/* also get the set of clauses that should be enforced by the scan */
allclauses = list_concat(list_copy(path->path.param_info->ppi_clauses), baserel->baserestrictinfo);
} else {
/* allclauses should just be the rel's restriction clauses */
allclauses = baserel->baserestrictinfo;
}
if (!u_sess->attr.attr_sql.enable_indexscan)
startup_cost += g_instance.cost_cxt.disable_cost;
/* we don't need to check enable_indexonlyscan; indxpath.c does that */
/*
* Call index-access-method-specific code to estimate the processing cost
* for scanning the index, as well as the selectivity of the index (ie,
* the fraction of main-table tuples we will have to retrieve) and its
* correlation to the main-table tuple order.
*/
OidFunctionCall8(index->amcostestimate,
PointerGetDatum(root),
PointerGetDatum(path),
Float8GetDatum(loop_count),
PointerGetDatum(&indexStartupCost),
PointerGetDatum(&indexTotalCost),
PointerGetDatum(&indexSelectivity),
PointerGetDatum(&indexCorrelation),
PointerGetDatum(&index_pages));
/*
* Save amcostestimate's results for possible use in bitmap scan planning.
* We don't bother to save indexStartupCost or indexCorrelation, because a
* bitmap scan doesn't care about either.
*/
path->indextotalcost = indexTotalCost;
path->indexselectivity = indexSelectivity;
/* all costs for touching index itself included here */
startup_cost += indexStartupCost;
run_cost += indexTotalCost - indexStartupCost;
/* estimate number of main-table tuples fetched */
tuples_fetched = clamp_row_est(indexSelectivity * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples));
/* fetch estimated page costs for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, &spc_random_page_cost, &spc_seq_page_cost);
/* ----------
* Estimate number of main-table pages fetched, and compute I/O cost.
*
* When the index ordering is uncorrelated with the table ordering,
* we use an approximation proposed by Mackert and Lohman (see
* index_pages_fetched() for details) to compute the number of pages
* fetched, and then charge spc_random_page_cost per page fetched.
*
* When the index ordering is exactly correlated with the table ordering
* (just after a CLUSTER, for example), the number of pages fetched should
* be exactly selectivity * table_size. What's more, all but the first
* will be sequential fetches, not the random fetches that occur in the
* uncorrelated case. So if the number of pages is more than 1, we
* ought to charge
* spc_random_page_cost + (pages_fetched - 1) * spc_seq_page_cost
* For partially-correlated indexes, we ought to charge somewhere between
* these two estimates. We currently interpolate linearly between the
* estimates based on the correlation squared (XXX is that appropriate?).
*
* If it's an index-only scan, then we will not need to fetch any heap
* pages for which the visibility map shows all tuples are visible.
* Hence, reduce the estimated number of heap fetches accordingly.
* We use the measured fraction of the entire heap that is all-visible,
* which might not be particularly relevant to the subset of the heap
* that this query will fetch; but it's not clear how to do better.
* ----------
*/
if (loop_count > 1) {
/*
* For repeated indexscans, the appropriate estimate for the
* uncorrelated case is to scale up the number of tuples fetched in
* the Mackert and Lohman formula by the number of scans, so that we
* estimate the number of pages fetched by all the scans; then
* pro-rate the costs for one scan. In this case we assume all the
* fetches are random accesses.
*/
pages_fetched = index_pages_fetched(
tuples_fetched * loop_count, (BlockNumber)baserel->pages, (double)index->pages, root, ispartitionedindex);
if (indexonly)
pages_fetched = ceil(pages_fetched * (1.0 - baserel->allvisfrac));
rand_heap_pages = pages_fetched;
max_IO_cost = (pages_fetched * spc_random_page_cost) / loop_count;
/*
* In the perfectly correlated case, the number of pages touched by
* each scan is selectivity * table_size, and we can use the Mackert
* and Lohman formula at the page level to estimate how much work is
* saved by caching across scans. We still assume all the fetches are
* random, though, which is an overestimate that's hard to correct for
* without double-counting the cache effects. (But in most cases
* where such a plan is actually interesting, only one page would get
* fetched per scan anyway, so it shouldn't matter much.)
*/
pages_fetched = ceil(indexSelectivity * (double)baserel->pages);
pages_fetched = index_pages_fetched(
pages_fetched * loop_count, (BlockNumber)baserel->pages, (double)index->pages, root, ispartitionedindex);
if (indexonly)
pages_fetched = ceil(pages_fetched * (1.0 - baserel->allvisfrac));
min_IO_cost = (pages_fetched * spc_random_page_cost) / loop_count;
} else {
/*
* Normal case: apply the Mackert and Lohman formula, and then
* interpolate between that and the correlation-derived result.
*/
pages_fetched = index_pages_fetched(
tuples_fetched, (BlockNumber)baserel->pages, (double)index->pages, root, ispartitionedindex);
if (indexonly)
pages_fetched = ceil(pages_fetched * (1.0 - baserel->allvisfrac));
rand_heap_pages = pages_fetched;
/* max_IO_cost is for the perfectly uncorrelated case (csquared=0) */
max_IO_cost = pages_fetched * spc_random_page_cost;
/* min_IO_cost is for the perfectly correlated case (csquared=1) */
pages_fetched = ceil(indexSelectivity * (double)baserel->pages);
if (indexonly)
pages_fetched = ceil(pages_fetched * (1.0 - baserel->allvisfrac));
if (pages_fetched > 0) {
min_IO_cost = spc_random_page_cost;
if (pages_fetched > 1)
min_IO_cost += (pages_fetched - 1) * spc_seq_page_cost;
} else {
min_IO_cost = 0;
}
}
if (partial_path) {
/*
* For index only scans compute workers based on number of index pages
* fetched; the number of heap pages we fetch might be so small as
* to effectively rule out parallelism, which we don't want to do.
*/
if (indexonly) {
rand_heap_pages = -1;
}
/*
* Estimate the number of parallel workers required to scan index. Use
* the number of heap pages computed considering heap fetches won't be
* sequential as for parallel scans the pages are accessed in random
* order.
*/
path->path.parallel_workers = compute_parallel_worker(baserel, rand_heap_pages,
index_pages, u_sess->attr.attr_sql.max_parallel_workers_per_gather);
/*
* Fall out if workers can't be assigned for parallel scan, because in
* such a case this path will be rejected. So there is no benefit in
* doing extra computation.
*/
if (path->path.parallel_workers <= 0) {
return;
}
path->path.parallel_aware = true;
}
/*
* Now interpolate based on estimated index order correlation to get total
* disk I/O cost for main table accesses.
*/
csquared = indexCorrelation * indexCorrelation;
run_cost += max_IO_cost + csquared * (min_IO_cost - max_IO_cost);
/*
* Estimate CPU costs per tuple.
*
* What we want here is cpu_tuple_cost plus the evaluation costs of any
* qual clauses that we have to evaluate as qpquals. We approximate that
* list as allclauses minus any clauses appearing in indexquals. (We
* assume that pointer equality is enough to recognize duplicate
* RestrictInfos.) This method neglects some considerations such as
* clauses that needn't be checked because they are implied by a partial
* index's predicate. It does not seem worth the cycles to try to factor
* those things in at this stage, even though createplan.c will take pains
* to remove such unnecessary clauses from the qpquals list if this path
* is selected for use.
*/
cost_qual_eval(&qpqual_cost, list_difference_ptr(allclauses, path->indexquals), root);
startup_cost += qpqual_cost.startup;
if (path->path.parent->orientation == REL_COL_ORIENTED)
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost / 10 + qpqual_cost.per_tuple;
else if (path->path.parent->orientation == REL_TIMESERIES_ORIENTED)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("Unsupported Using Index FOR TIMESERIES.")));
else
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qpqual_cost.per_tuple;
cpu_run_cost += cpu_per_tuple * tuples_fetched;
/* Adjust costing for parallelism, if used. */
if (path->path.parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(&path->path);
path->path.rows = clamp_row_est(path->path.rows / parallel_divisor);
/* The CPU cost is divided among all the workers. */
cpu_run_cost /= parallel_divisor;
}
run_cost += cpu_run_cost;
path->path.startup_cost = startup_cost;
path->path.total_cost = startup_cost + run_cost;
path->path.stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_indexscan)
path->path.total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* index_pages_fetched
* Estimate the number of pages actually fetched after accounting for
* cache effects.
*
* We use an approximation proposed by Mackert and Lohman, "Index Scans
* Using a Finite LRU Buffer: A Validated I/O Model", ACM Transactions
* on Database Systems, Vol. 14, No. 3, September 1989, Pages 401-424.
* The Mackert and Lohman approximation is that the number of pages
* fetched is
* PF =
* min(2TNs/(2T+Ns), T) when T <= b
* 2TNs/(2T+Ns) when T > b and Ns <= 2Tb/(2T-b)
* b + (Ns - 2Tb/(2T-b))*(T-b)/T when T > b and Ns > 2Tb/(2T-b)
* where
* T = # pages in table
* N = # tuples in table
* s = selectivity = fraction of table to be scanned
* b = # buffer pages available (we include kernel space here)
*
* We assume that g_instance.cost_cxt.effective_cache_size is the total number of buffer pages
* available for the whole query, and pro-rate that space across all the
* tables in the query and the index currently under consideration. (This
* ignores space needed for other indexes used by the query, but since we
* don't know which indexes will get used, we can't estimate that very well;
* and in any case counting all the tables may well be an overestimate, since
* depending on the join plan not all the tables may be scanned concurrently.)
*
* The product Ns is the number of tuples fetched; we pass in that
* product rather than calculating it here. "pages" is the number of pages
* in the object under consideration (either an index or a table).
* "index_pages" is the amount to add to the total table space, which was
* computed for us by query_planner.
*
* Caller is expected to have ensured that tuples_fetched is greater than zero
* and rounded to integer (see clamp_row_est). The result will likewise be
* greater than zero and integral.
*
* add an input parameter to indicate if it is for partitioned index. because the
* method of calculating the base stat info for partitioned index cannot fulfill
* the logic for ordinary table., so we have to deal with it specially.
*/
double index_pages_fetched(
double tuples_fetched, BlockNumber pages, double index_pages, PlannerInfo* root, bool ispartitionedindex)
{
double pages_fetched;
double total_pages;
double T, b;
/* T is # pages in table, but don't allow it to be zero */
T = (pages > 1) ? (double)pages : 1.0;
/* Compute number of pages assumed to be competing for cache space */
total_pages = root->total_table_pages + index_pages;
total_pages = Max(total_pages, 1.0);
/*
* it is special to estimate the number of pages actually fetched.
* it look likes illegal, but we have to do this because the method
* of calculating the base stat info for partitioned index cannot
* fulfill the logic for ordinary table
*/
if (ispartitionedindex) {
T = (T > total_pages ? T : total_pages);
} else {
AssertEreport(total_pages >= T,
MOD_OPT,
"The number of pages in table is larger than total_pages"
"when estimating the number of pages actually fetched.");
}
/* b is pro-rated share of u_sess->attr.attr_sql.effective_cache_size */
b = (double)u_sess->attr.attr_sql.effective_cache_size * T / total_pages;
/* force it positive and integral */
if (b <= 1.0) {
b = 1.0;
} else {
b = ceil(b);
}
/* This part is the Mackert and Lohman formula */
if (T <= b) {
pages_fetched = (2.0 * T * tuples_fetched) / (2.0 * T + tuples_fetched);
if (pages_fetched >= T) {
pages_fetched = T;
} else {
pages_fetched = ceil(pages_fetched);
}
} else {
double lim;
lim = (2.0 * T * b) / (2.0 * T - b);
if (tuples_fetched <= lim) {
pages_fetched = (2.0 * T * tuples_fetched) / (2.0 * T + tuples_fetched);
} else {
pages_fetched = b + (tuples_fetched - lim) * (T - b) / T;
}
pages_fetched = ceil(pages_fetched);
}
return pages_fetched;
}
/*
* get_indexpath_pages
* Determine the total size of the indexes used in a bitmap index path.
*
* Note: if the same index is used more than once in a bitmap tree, we will
* count it multiple times, which perhaps is the wrong thing ... but it's
* not completely clear, and detecting duplicates is difficult, so ignore it
* for now.
*/
static double get_indexpath_pages(Path* bitmapqual)
{
double result = 0;
ListCell* l = NULL;
if (IsA(bitmapqual, BitmapAndPath)) {
BitmapAndPath* apath = (BitmapAndPath*)bitmapqual;
foreach (l, apath->bitmapquals) {
result += get_indexpath_pages((Path*)lfirst(l));
}
} else if (IsA(bitmapqual, BitmapOrPath)) {
BitmapOrPath* opath = (BitmapOrPath*)bitmapqual;
foreach (l, opath->bitmapquals) {
result += get_indexpath_pages((Path*)lfirst(l));
}
} else if (IsA(bitmapqual, IndexPath)) {
IndexPath* ipath = (IndexPath*)bitmapqual;
result = (double)ipath->indexinfo->pages;
} else {
ereport(ERROR,
(errmodule(MOD_OPT),
errcode(ERRCODE_UNRECOGNIZED_NODE_TYPE),
errmsg("unrecognized node type of a bitmap index path when get pages: %d", nodeTag(bitmapqual))));
}
return result;
}
/*
* cost_bitmap_heap_scan
* Determines and returns the cost of scanning a relation using a bitmap
* index-then-heap plan.
*
* 'baserel' is the relation to be scanned
* 'param_info' is the ParamPathInfo if this is a parameterized path, else NULL
* 'bitmapqual' is a tree of IndexPaths, BitmapAndPaths, and BitmapOrPaths
* 'loop_count' is the number of repetitions of the indexscan to factor into
* estimates of caching behavior
*
* Note: the component IndexPaths in bitmapqual should have been costed
* using the same loop_count.
*/
void cost_bitmap_heap_scan(
Path* path, PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info, Path* bitmapqual, double loop_count)
{
Cost startup_cost = 0;
Cost run_cost = 0;
Cost indexTotalCost;
QualCost qpqual_cost;
Cost cpu_per_tuple = 0.0;
Cost cost_per_page;
Cost cpu_run_cost;
double tuples_fetched;
double pages_fetched;
double spc_seq_page_cost, spc_random_page_cost;
double T;
bool ispartitionedindex = path->parent->isPartitionedTable;
bool partition_index_unusable = false;
bool containGlobalOrLocalIndex = false;
/* Should only be applied to base relations */
AssertEreport(IsA(baserel, RelOptInfo),
MOD_OPT,
"The nodeTag of baserel is not T_RelOptInfo"
"when determining the cost of scanning a relation using a bitmap index-then-heap plan.");
AssertEreport(baserel->relid > 0,
MOD_OPT,
"The relid is invalid when determining the cost of scanning a relation"
"using a bitmap index-then-heap plan.");
AssertEreport(baserel->rtekind == RTE_RELATION,
MOD_OPT,
"Only base relation can be supported when determining the cost of scanning a relation"
"using a bitmap index-then-heap plan.");
/* Mark the path with the correct row estimate */
set_rel_path_rows(path, baserel, param_info);
/*
* Support partiton index unusable.
* Here not support bitmap index unusable.If the bitmap path contains unusable index paths, set enable_bitmapscan to
* off. So it will go partition full/partial unusable index scan ,if index path is selected.
*/
if (ispartitionedindex) {
if (!check_bitmap_heap_path_index_unusable(bitmapqual, baserel))
partition_index_unusable = true;
/* If the bitmap path contains Global partition index OR local partition index, set enable_bitmapscan to off */
if (CheckBitmapHeapPathContainGlobalOrLocal(bitmapqual)) {
containGlobalOrLocalIndex = true;
}
}
if (!u_sess->attr.attr_sql.enable_bitmapscan || partition_index_unusable ||
containGlobalOrLocalIndex) {
startup_cost += g_instance.cost_cxt.disable_cost;
}
pages_fetched = compute_bitmap_pages(root, baserel, bitmapqual, loop_count,
&indexTotalCost, &tuples_fetched, ispartitionedindex);
startup_cost += indexTotalCost;
/* Fetch estimated page costs for tablespace containing table. */
get_tablespace_page_costs(baserel->reltablespace, &spc_random_page_cost, &spc_seq_page_cost);
T = (baserel->pages > 1) ? (double)baserel->pages : 1.0;
/*
* For small numbers of pages we should charge spc_random_page_cost
* apiece, while if nearly all the table's pages are being read, it's more
* appropriate to charge spc_seq_page_cost apiece. The effect is
* nonlinear, too. For lack of a better idea, interpolate like this to
* determine the cost per page.
*/
if (pages_fetched >= 2.0)
cost_per_page = spc_random_page_cost - (spc_random_page_cost - spc_seq_page_cost) * sqrt(pages_fetched / T);
else
cost_per_page = spc_random_page_cost;
run_cost += pages_fetched * cost_per_page;
/*
* Estimate CPU costs per tuple.
*
* Often the indexquals don't need to be rechecked at each tuple ... but
* not always, especially not if there are enough tuples involved that the
* bitmaps become lossy. For the moment, just assume they will be
* rechecked always. This means we charge the full freight for all the
* scan clauses.
*/
get_restriction_qual_cost(root, baserel, param_info, &qpqual_cost);
startup_cost += qpqual_cost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qpqual_cost.per_tuple;
cpu_run_cost = cpu_per_tuple * tuples_fetched;
/* Adjust costing for parallelism, if used. */
if (path->parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(path);
/* The CPU cost is divided among all the workers. */
cpu_run_cost /= parallel_divisor;
path->rows = clamp_row_est(path->rows / parallel_divisor);
}
run_cost += cpu_run_cost;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_bitmapscan || partition_index_unusable)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* cost_bitmap_tree_node
* Extract cost and selectivity from a bitmap tree node (index/and/or)
*/
void cost_bitmap_tree_node(Path* path, Cost* cost, Selectivity* selec)
{
if (IsA(path, IndexPath)) {
*cost = ((IndexPath*)path)->indextotalcost;
*selec = ((IndexPath*)path)->indexselectivity;
/*
* Charge a small amount per retrieved tuple to reflect the costs of
* manipulating the bitmap. This is mostly to make sure that a bitmap
* scan doesn't look to be the same cost as an indexscan to retrieve a
* single tuple.
*/
*cost += 0.1 * u_sess->attr.attr_sql.cpu_operator_cost * PATH_LOCAL_ROWS(path);
} else if (IsA(path, BitmapAndPath)) {
*cost = path->total_cost;
*selec = ((BitmapAndPath*)path)->bitmapselectivity;
} else if (IsA(path, BitmapOrPath)) {
*cost = path->total_cost;
*selec = ((BitmapOrPath*)path)->bitmapselectivity;
} else {
ereport(ERROR,
(errmodule(MOD_OPT),
errcode(ERRCODE_UNRECOGNIZED_NODE_TYPE),
errmsg("unrecognized node type when extract cost and selectivity from a bitmap tree node: %d",
nodeTag(path))));
*cost = *selec = 0; /* keep compiler quiet */
}
}
/*
* cost_bitmap_and_node
* Estimate the cost of a BitmapAnd node
*
* Note that this considers only the costs of index scanning and bitmap
* creation, not the eventual heap access. In that sense the object isn't
* truly a Path, but it has enough path-like properties (costs in particular)
* to warrant treating it as one. We don't bother to set the path rows field,
* however.
*/
void cost_bitmap_and_node(BitmapAndPath* path, PlannerInfo* root)
{
Cost totalCost;
Selectivity selec;
ListCell* l = NULL;
/*
* We estimate AND selectivity on the assumption that the inputs are
* independent. This is probably often wrong, but we don't have the info
* to do better.
*
* The runtime cost of the BitmapAnd itself is estimated at 100x
* cpu_operator_cost for each tbm_intersect needed. Probably too small,
* definitely too simplistic?
*/
totalCost = 0.0;
selec = 1.0;
foreach (l, path->bitmapquals) {
Path* subpath = (Path*)lfirst(l);
Cost subCost;
Selectivity subselec;
cost_bitmap_tree_node(subpath, &subCost, &subselec);
selec *= subselec;
totalCost += subCost;
if (l != list_head(path->bitmapquals))
totalCost += 100.0 * u_sess->attr.attr_sql.cpu_operator_cost;
}
path->bitmapselectivity = selec;
set_path_rows(&path->path, 0); /* per above, not used */
path->path.startup_cost = totalCost;
path->path.total_cost = totalCost;
path->path.stream_cost = 0;
}
/*
* cost_bitmap_or_node
* Estimate the cost of a BitmapOr node
*
* See comments for cost_bitmap_and_node.
*/
void cost_bitmap_or_node(BitmapOrPath* path, PlannerInfo* root)
{
Cost totalCost;
Selectivity selec;
ListCell* l = NULL;
/*
* We estimate OR selectivity on the assumption that the inputs are
* non-overlapping, since that's often the case in "x IN (list)" type
* situations. Of course, we clamp to 1.0 at the end.
*
* The runtime cost of the BitmapOr itself is estimated at 100x
* cpu_operator_cost for each tbm_union needed. Probably too small,
* definitely too simplistic? We are aware that the tbm_unions are
* optimized out when the inputs are BitmapIndexScans.
*/
totalCost = 0.0;
selec = 0.0;
foreach (l, path->bitmapquals) {
Path* subpath = (Path*)lfirst(l);
Cost subCost;
Selectivity subselec;
cost_bitmap_tree_node(subpath, &subCost, &subselec);
selec += subselec;
totalCost += subCost;
if (l != list_head(path->bitmapquals) && !IsA(subpath, IndexPath))
totalCost += 100.0 * u_sess->attr.attr_sql.cpu_operator_cost;
}
path->bitmapselectivity = Min(selec, 1.0);
set_path_rows(&path->path, 0); /* per above, not used */
path->path.startup_cost = totalCost;
path->path.total_cost = totalCost;
path->path.stream_cost = 0;
}
/*
* cost_tidscan
* Determines and returns the cost of scanning a relation using TIDs.
*/
void cost_tidscan(Path* path, PlannerInfo* root, RelOptInfo* baserel, List* tidquals)
{
Cost startup_cost = 0;
Cost run_cost = 0;
bool isCurrentOf = false;
Cost cpu_per_tuple = 0.0;
QualCost tid_qual_cost;
int ntuples;
ListCell* l = NULL;
double spc_random_page_cost;
/* Should only be applied to base relations */
AssertEreport(baserel->relid > 0,
MOD_OPT,
"The relid is invalid when determining the cost of scanning a relation using TIDs.");
AssertEreport(baserel->rtekind == RTE_RELATION,
MOD_OPT,
"Only base relation can be supported"
"when determining the cost of scanning a relation using TIDs.");
/* For now, tidscans are never parameterized */
set_rel_path_rows(path, baserel, NULL);
/* Count how many tuples we expect to retrieve */
ntuples = 0;
foreach (l, tidquals) {
if (IsA(lfirst(l), ScalarArrayOpExpr)) {
/* Each element of the array yields 1 tuple */
ScalarArrayOpExpr* saop = (ScalarArrayOpExpr*)lfirst(l);
Node* arraynode = (Node*)lsecond(saop->args);
ntuples += estimate_array_length(arraynode);
} else if (IsA(lfirst(l), CurrentOfExpr)) {
/* CURRENT OF yields 1 tuple */
isCurrentOf = true;
ntuples++;
} else {
/* It's just CTID = something, count 1 tuple */
ntuples++;
}
}
/*
* We must force TID scan for WHERE CURRENT OF, because only nodeTidscan.c
* understands how to do it correctly. Therefore, honor u_sess->attr.attr_sql.enable_tidscan
* only when CURRENT OF isn't present. Also note that cost_qual_eval
* counts a CurrentOfExpr as having startup cost g_instance.cost_cxt.disable_cost, which we
* subtract off here; that's to prevent other plan types such as seqscan
* from winning.
*/
if (isCurrentOf) {
AssertEreport(baserel->baserestrictcost.startup >= g_instance.cost_cxt.disable_cost,
MOD_OPT,
"The one-time cost of base relation is less than disable_cost"
"when determining the cost of scanning a relation using TIDs.");
startup_cost -= g_instance.cost_cxt.disable_cost;
} else if (!u_sess->attr.attr_sql.enable_tidscan)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* The TID qual expressions will be computed once, any other baserestrict
* quals once per retrieved tuple.
*/
cost_qual_eval(&tid_qual_cost, tidquals, root);
/* fetch estimated page cost for tablespace containing table */
get_tablespace_page_costs(baserel->reltablespace, &spc_random_page_cost, NULL);
/* disk costs --- assume each tuple on a different page */
run_cost += spc_random_page_cost * ntuples;
/* CPU costs */
startup_cost += baserel->baserestrictcost.startup + tid_qual_cost.per_tuple;
cpu_per_tuple =
u_sess->attr.attr_sql.cpu_tuple_cost + baserel->baserestrictcost.per_tuple - tid_qual_cost.per_tuple;
run_cost += cpu_per_tuple * ntuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_tidscan && !isCurrentOf)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* cost_subqueryscan
* Determines and returns the cost of scanning a subquery RTE.
*
* 'baserel' is the relation to be scanned
* 'param_info' is the ParamPathInfo if this is a parameterized path, else NULL
*/
void cost_subqueryscan(Path* path, PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info)
{
Cost startup_cost;
Cost run_cost;
QualCost qpqual_cost;
Cost cpu_per_tuple = 0.0;
/* Should only be applied to base relations that are subqueries */
AssertEreport(
baserel->relid > 0, MOD_OPT, "The relid is invalid when determining the cost of scanning a subquery RTE.");
AssertEreport(baserel->rtekind == RTE_SUBQUERY,
MOD_OPT,
"Only subquery in FROM clause can be supported"
"when determining the cost of scanning a subquery RTE.");
/* Mark the path with the correct row estimate */
set_rel_path_rows(path, baserel, param_info);
/*
* Cost of path is cost of evaluating the subplan, plus cost of evaluating
* any restriction clauses that will be attached to the SubqueryScan node,
* plus cpu_tuple_cost to account for selection and projection overhead.
*/
path->startup_cost = baserel->subplan->startup_cost;
path->total_cost = baserel->subplan->total_cost;
get_restriction_qual_cost(root, baserel, param_info, &qpqual_cost);
startup_cost = qpqual_cost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qpqual_cost.per_tuple;
run_cost = cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples);
path->startup_cost += startup_cost;
path->total_cost += startup_cost + run_cost;
path->stream_cost = get_subqueryscan_stream_cost(baserel->subplan);
ereport(DEBUG2,
(errmodule(MOD_OPT_SUBPLAN),
errmsg("subqueryscan stream_cost: %lf, startup_cost: %lf, total_cost: %lf",
path->stream_cost,
path->startup_cost,
path->total_cost)));
}
/*
* cost_functionscan
* Determines and returns the cost of scanning a function RTE.
*/
void cost_functionscan(Path* path, PlannerInfo* root, RelOptInfo* baserel)
{
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_per_tuple = 0.0;
RangeTblEntry* rte = NULL;
QualCost exprcost;
/* Should only be applied to base relations that are functions */
AssertEreport(
baserel->relid > 0, MOD_OPT, "The relid is invalid when determining the cost of scanning a function RTE.");
rte = planner_rt_fetch(baserel->relid, root);
AssertEreport(rte->rtekind == RTE_FUNCTION,
MOD_OPT,
"Only function in FROM clause can be supported"
"when determining the cost of scanning a function RTE.");
/* functionscans are never parameterized */
set_rel_path_rows(path, baserel, NULL);
/*
* Estimate costs of executing the function expression.
*
* Currently, nodeFunctionscan.c always executes the function to
* completion before returning any rows, and caches the results in a
* tuplestore. So the function eval cost is all startup cost, and per-row
* costs are minimal.
*
* XXX in principle we ought to charge tuplestore spill costs if the
* number of rows is large. However, given how phony our rowcount
* estimates for functions tend to be, there's not a lot of point in that
* refinement right now.
*/
cost_qual_eval_node(&exprcost, rte->funcexpr, root);
startup_cost += exprcost.startup + exprcost.per_tuple;
/* Add scanning CPU costs */
startup_cost += baserel->baserestrictcost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + baserel->baserestrictcost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples);
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
}
/*
* cost_valuesscan
* Determines and returns the cost of scanning a VALUES RTE.
*/
void cost_valuesscan(Path* path, PlannerInfo* root, RelOptInfo* baserel)
{
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_per_tuple = 0.0;
/* Should only be applied to base relations that are values lists */
AssertEreport(
baserel->relid > 0, MOD_OPT, "The relid is invalid when determining the cost of scanning a VALUES RTE.");
AssertEreport(baserel->rtekind == RTE_VALUES,
MOD_OPT,
"Only VALUES list can be supported when determining the cost of scanning a VALUES RTE.");
/* valuesscans are never parameterized */
set_rel_path_rows(path, baserel, NULL);
/*
* For now, estimate list evaluation cost at one operator eval per list
* (probably pretty bogus, but is it worth being smarter?)
*/
cpu_per_tuple = u_sess->attr.attr_sql.cpu_operator_cost;
/* Add scanning CPU costs */
startup_cost += baserel->baserestrictcost.startup;
cpu_per_tuple += u_sess->attr.attr_sql.cpu_tuple_cost + baserel->baserestrictcost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples);
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
}
/*
* cost_ctescan
* Determines and returns the cost of scanning a CTE RTE.
*
* Note: this is used for both self-reference and regular CTEs; the
* possible cost differences are below the threshold of what we could
* estimate accurately anyway. Note that the costs of evaluating the
* referenced CTE query are added into the final plan as initplan costs,
* and should NOT be counted here.
*/
void cost_ctescan(Path* path, PlannerInfo* root, RelOptInfo* baserel)
{
Cost startup_cost = 0;
Cost run_cost = 0;
Cost cpu_per_tuple = 0.0;
/* Should only be applied to base relations that are CTEs */
AssertEreport(baserel->relid > 0, MOD_OPT, "The relid is invalid when determining the cost of scanning a CTE RTE.");
AssertEreport(baserel->rtekind == RTE_CTE,
MOD_OPT,
"Only common table expr can be supported when determining the cost of scanning a CTE RTE.");
/* ctescans are never parameterized */
set_rel_path_rows(path, baserel, NULL);
/* Charge one CPU tuple cost per row for tuplestore manipulation */
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost;
/* Add scanning CPU costs */
startup_cost += baserel->baserestrictcost.startup;
cpu_per_tuple += u_sess->attr.attr_sql.cpu_tuple_cost + baserel->baserestrictcost.per_tuple;
run_cost += cpu_per_tuple * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples);
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
}
/*
* cost_recursive_union
* Determines and returns the cost of performing a recursive union,
* and also the estimated output size.
*
* We are given Plans for the nonrecursive and recursive terms.
*
* Note that the arguments and output are Plans, not Paths as in most of
* the rest of this module. That's because we don't bother setting up a
* Path representation for recursive union --- we have only one way to do it.
*/
void cost_recursive_union(Plan* runion, Plan* nrterm, Plan* rterm)
{
Cost startup_cost;
Cost total_cost;
double total_rows;
double total_global_rows;
/* We probably have decent estimates for the non-recursive term */
startup_cost = nrterm->startup_cost;
total_cost = nrterm->total_cost;
total_rows = PLAN_LOCAL_ROWS(nrterm);
total_global_rows = nrterm->plan_rows;
/*
* We arbitrarily assume that about 10 recursive iterations will be
* needed, and that we've managed to get a good fix on the cost and output
* size of each one of them. These are mighty shaky assumptions but it's
* hard to see how to do better.
*/
total_cost += 10 * rterm->total_cost;
total_rows += 10 * PLAN_LOCAL_ROWS(rterm);
total_global_rows += 10 * rterm->plan_rows;
/*
* Also charge cpu_tuple_cost per row to account for the costs of
* manipulating the tuplestores. (We don't worry about possible
* spill-to-disk costs.)
*/
total_cost += u_sess->attr.attr_sql.cpu_tuple_cost * total_rows;
runion->startup_cost = startup_cost;
runion->total_cost = total_cost;
set_plan_rows(runion, total_global_rows, nrterm->multiple);
runion->plan_width = Max(nrterm->plan_width, rterm->plan_width);
}
/*
* cost_sort
* Determines and returns the cost of sorting a relation, including
* the cost of reading the input data.
*
* If the total volume of data to sort is less than sort_mem, we will do
* an in-memory sort, which requires no I/O and about t*log2(t) tuple
* comparisons for t tuples.
*
* If the total volume exceeds sort_mem, we switch to a tape-style merge
* algorithm. There will still be about t*log2(t) tuple comparisons in
* total, but we will also need to write and read each tuple once per
* merge pass. We expect about ceil(logM(r)) merge passes where r is the
* number of initial runs formed and M is the merge order used by tuplesort.c.
* Since the average initial run should be about twice sort_mem, we have
* disk traffic = 2 * relsize * ceil(logM(p / (2*sort_mem)))
* cpu = comparison_cost * t * log2(t)
*
* If the sort is bounded (i.e., only the first k result tuples are needed)
* and k tuples can fit into sort_mem, we use a heap method that keeps only
* k tuples in the heap; this will require about t*log2(k) tuple comparisons.
*
* The disk traffic is assumed to be 3/4ths sequential and 1/4th random
* accesses (XXX can't we refine that guess?)
*
* By default, we charge two operator evals per tuple comparison, which should
* be in the right ballpark in most cases. The caller can tweak this by
* specifying nonzero comparison_cost; typically that's used for any extra
* work that has to be done to prepare the inputs to the comparison operators.
*
* 'pathkeys' is a list of sort keys
* 'input_cost' is the total cost for reading the input data
* 'tuples' is the number of tuples in the relation
* 'width' is the average tuple width in bytes
* 'comparison_cost' is the extra cost per comparison, if any
* 'sort_mem' is the number of kilobytes of work memory allowed for the sort
* 'limit_tuples' is the bound on the number of output tuples; -1 if no bound
* 'mem_info' is operator max and min info used by memory control module
*
* NOTE: some callers currently pass NIL for pathkeys because they
* can't conveniently supply the sort keys. Since this routine doesn't
* currently do anything with pathkeys anyway, that doesn't matter...
* but if it ever does, it should react gracefully to lack of key data.
* (Actually, the thing we'd most likely be interested in is just the number
* of sort keys, which all callers *could* supply.)
*/
void cost_sort(Path* path, List* pathkeys, Cost input_cost, double tuples, int width, Cost comparison_cost,
int sort_mem, double limit_tuples, bool col_store, int dop, OpMemInfo* mem_info, bool index_sort)
{
Cost startup_cost = input_cost;
Cost run_cost = 0;
double input_bytes = relation_byte_size(tuples, width, col_store, true, true, index_sort) / SET_DOP(dop);
double output_bytes;
double output_tuples;
long sort_mem_bytes = sort_mem * 1024L / SET_DOP(dop);
dop = SET_DOP(dop);
if (!u_sess->attr.attr_sql.enable_sort)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* We want to be sure the cost of a sort is never estimated as zero, even
* if passed-in tuple count is zero. Besides, mustn't do log(0)...
*/
if (tuples < 2.0) {
tuples = 2.0;
}
/* Include the default cost-per-comparison */
comparison_cost += 2.0 * u_sess->attr.attr_sql.cpu_operator_cost;
/* Do we have a useful LIMIT? */
if (limit_tuples > 0 && limit_tuples < tuples) {
output_tuples = limit_tuples;
output_bytes = relation_byte_size(output_tuples, width, col_store, true, true, index_sort);
} else {
output_tuples = tuples;
output_bytes = input_bytes;
}
if (output_bytes > sort_mem_bytes) {
/*
* CPU costs
*
* Assume about N log2 N comparisons
*/
startup_cost += comparison_cost * tuples * LOG2(tuples);
/* Disk costs */
startup_cost += compute_sort_disk_cost(input_bytes, sort_mem_bytes);
} else {
if (tuples > 2 * output_tuples || input_bytes > sort_mem_bytes) {
/*
* We'll use a bounded heap-sort keeping just K tuples in memory, for
* a total number of tuple comparisons of N log2 K; but the constant
* factor is a bit higher than for quicksort. Tweak it so that the
* cost curve is continuous at the crossover point.
*/
startup_cost += comparison_cost * tuples * LOG2(2.0 * output_tuples);
} else {
/* We'll use plain quicksort on all the input tuples */
startup_cost += comparison_cost * tuples * LOG2(tuples);
}
}
if (mem_info != NULL) {
mem_info->opMem = u_sess->opt_cxt.op_work_mem;
mem_info->maxMem = output_bytes / 1024L * dop;
mem_info->minMem = mem_info->maxMem / SORT_MAX_DISK_SIZE;
mem_info->regressCost = compute_sort_disk_cost(input_bytes, mem_info->minMem);
/* Special case if array larger than 1G, so we must spill to disk */
if (output_tuples > (MaxAllocSize / TUPLE_OVERHEAD(true) * dop)) {
mem_info->maxMem = STATEMENT_MIN_MEM * 1024L * dop;
mem_info->minMem = Min(mem_info->maxMem, mem_info->minMem);
}
}
/*
* Also charge a small amount (arbitrarily set equal to operator cost) per
* extracted tuple. We don't charge cpu_tuple_cost because a Sort node
* doesn't do qual-checking or projection, so it has less overhead than
* most plan nodes. Note it's correct to use tuples not output_tuples
* here --- the upper LIMIT will pro-rate the run cost so we'd be double
* counting the LIMIT otherwise.
*/
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * tuples;
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
path->stream_cost = 0;
if (!u_sess->attr.attr_sql.enable_sort)
path->total_cost *=
(g_instance.cost_cxt.disable_cost_enlarge_factor * g_instance.cost_cxt.disable_cost_enlarge_factor);
}
/*
* append_nonpartial_cost
* Estimate the cost of the non-partial paths in a Parallel Append.
* The non-partial paths are assumed to be the first "numpaths" paths
* from the subpaths list, and to be in order of decreasing cost.
*/
static Cost append_nonpartial_cost(List *subpaths, int numpaths, int parallel_workers)
{
Cost *costarr = NULL;
int arrlen;
ListCell *l = NULL;
ListCell *cell = NULL;
int i;
int path_index;
int min_index;
int max_index;
if (numpaths == 0) {
return 0;
}
/*
* Array length is number of workers or number of relevants paths,
* whichever is less.
*/
arrlen = Min(parallel_workers, numpaths);
costarr = (Cost *) palloc(sizeof(Cost) * arrlen);
/* The first few paths will each be claimed by a different worker. */
path_index = 0;
foreach(cell, subpaths) {
Path *subpath = (Path *) lfirst(cell);
if (path_index == arrlen)
break;
costarr[path_index++] = subpath->total_cost;
}
/*
* Since subpaths are sorted by decreasing cost, the last one will have
* the minimum cost.
*/
min_index = arrlen - 1;
/*
* For each of the remaining subpaths, add its cost to the array element
* with minimum cost.
*/
for_each_cell(l, cell) {
Path *subpath = (Path *) lfirst(l);
int i;
/* Consider only the non-partial paths */
if (path_index++ == numpaths)
break;
costarr[min_index] += subpath->total_cost;
/* Update the new min cost array index */
for (min_index = i = 0; i < arrlen; i++) {
if (costarr[i] < costarr[min_index])
min_index = i;
}
}
/* Return the highest cost from the array */
for (max_index = i = 0; i < arrlen; i++) {
if (costarr[i] > costarr[max_index])
max_index = i;
}
return costarr[max_index];
}
/*
* cost_append
* Determines and returns the cost of an Append node.
*
* We charge nothing extra for the Append itself, which perhaps is too
* optimistic, but since it doesn't do any selection or projection, it is a
* pretty cheap node.
*/
void cost_append(AppendPath *apath)
{
ListCell *l = NULL;
apath->path.startup_cost = 0;
apath->path.total_cost = 0;
if (apath->subpaths == NIL)
return;
if (!apath->path.parallel_aware) {
Path *subpath = (Path *) linitial(apath->subpaths);
/*
* Startup cost of non-parallel-aware Append is the startup cost of
* first subpath.
*/
apath->path.startup_cost = subpath->startup_cost;
/* Compute rows and costs as sums of subplan rows and costs. */
foreach(l, apath->subpaths) {
Path *subpath = (Path *) lfirst(l);
apath->path.rows += subpath->rows;
apath->path.total_cost += subpath->total_cost;
apath->path.stream_cost += subpath->stream_cost;
}
} else { /* parallel-aware */
int i = 0;
double parallel_divisor = get_parallel_divisor(&apath->path);
/* Calculate startup cost. */
foreach(l, apath->subpaths) {
Path *subpath = (Path *) lfirst(l);
/*
* Append will start returning tuples when the child node having
* lowest startup cost is done setting up. We consider only the
* first few subplans that immediately get a worker assigned.
*/
if (i == 0) {
apath->path.startup_cost = subpath->startup_cost;
} else if (i < apath->path.parallel_workers) {
apath->path.startup_cost = Min(apath->path.startup_cost, subpath->startup_cost);
}
/*
* Apply parallel divisor to subpaths. Scale the number of rows
* for each partial subpath based on the ratio of the parallel
* divisor originally used for the subpath to the one we adopted.
* Also add the cost of partial paths to the total cost, but
* ignore non-partial paths for now.
*/
if (i < apath->first_partial_path) {
apath->path.rows += subpath->rows / parallel_divisor;
} else {
double subpath_parallel_divisor = get_parallel_divisor(subpath);
apath->path.rows += subpath->rows * (subpath_parallel_divisor / parallel_divisor);
apath->path.total_cost += subpath->total_cost;
}
apath->path.rows = clamp_row_est(apath->path.rows);
apath->path.stream_cost += subpath->stream_cost;
i++;
}
/* Add cost for non-partial subpaths. */
apath->path.total_cost += append_nonpartial_cost(apath->subpaths, apath->first_partial_path,
apath->path.parallel_workers);
}
}
/*
* compute_sort_disk_cost
* compute disk spill cost of sort operator
*
* Parameters:
* @in input_bytes: bytes of input relation
* @in sort_mem_bytes: work mem of sort operator
*
* Returns: estimated disk cost
*/
double compute_sort_disk_cost(double input_bytes, double sort_mem_bytes)
{
/*
* We'll have to use a disk-based sort of all the tuples
*/
double npages = ceil(input_bytes / BLCKSZ);
double nruns = (input_bytes / sort_mem_bytes) * 0.5;
double mergeorder = tuplesort_merge_order(sort_mem_bytes);
double log_runs;
double npageaccesses;
/* Compute logM(r) as log(r) / log(M) */
if (nruns > mergeorder) {
log_runs = ceil(log(nruns) / log(mergeorder));
} else {
log_runs = 1.0;
}
npageaccesses = 2.0 * npages * log_runs;
/* Assume 3/4ths of accesses are sequential, 1/4th are not */
return npageaccesses * (u_sess->attr.attr_sql.seq_page_cost * 0.75 + u_sess->attr.attr_sql.random_page_cost * 0.25);
}
/*
* cost_merge_append
* Determines and returns the cost of a MergeAppend node.
*
* MergeAppend merges several pre-sorted input streams, using a heap that
* at any given instant holds the next tuple from each stream. If there
* are N streams, we need about N*log2(N) tuple comparisons to construct
* the heap at startup, and then for each output tuple, about log2(N)
* comparisons to delete the top heap entry and another log2(N) comparisons
* to insert its successor from the same stream.
*
* (The effective value of N will drop once some of the input streams are
* exhausted, but it seems unlikely to be worth trying to account for that.)
*
* The heap is never spilled to disk, since we assume N is not very large.
* So this is much simpler than cost_sort.
*
* As in cost_sort, we charge two operator evals per tuple comparison.
*
* 'pathkeys' is a list of sort keys
* 'n_streams' is the number of input streams
* 'input_startup_cost' is the sum of the input streams' startup costs
* 'input_total_cost' is the sum of the input streams' total costs
* 'tuples' is the number of tuples in all the streams
*/
void cost_merge_append(Path* path, PlannerInfo* root, List* pathkeys, int n_streams, Cost input_startup_cost,
Cost input_total_cost, double tuples)
{
Cost startup_cost = 0;
Cost run_cost = 0;
Cost comparison_cost;
double N;
double logN;
/*
* Avoid log(0)...
*/
N = (n_streams < 2) ? 2.0 : (double)n_streams;
logN = LOG2(N);
/* Assumed cost per tuple comparison */
comparison_cost = 2.0 * u_sess->attr.attr_sql.cpu_operator_cost;
/* Heap creation cost */
startup_cost += comparison_cost * N * logN;
/* Per-tuple heap maintenance cost */
run_cost += tuples * comparison_cost * 2.0 * logN;
/*
* Also charge a small amount (arbitrarily set equal to operator cost) per
* extracted tuple. We don't charge cpu_tuple_cost because a MergeAppend
* node doesn't do qual-checking or projection, so it has less overhead
* than most plan nodes.
*/
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * tuples;
path->startup_cost = startup_cost + input_startup_cost;
path->total_cost = startup_cost + run_cost + input_total_cost;
}
/*
* cost_material
* Determines and returns the cost of materializing a relation, including
* the cost of reading the input data.
*
* If the total volume of data to materialize exceeds work_mem, we will need
* to write it to disk, so the cost is much higher in that case.
*
* Note that here we are estimating the costs for the first scan of the
* relation, so the materialization is all overhead --- any savings will
* occur only on rescan, which is estimated in cost_rescan.
*/
void cost_material(Path* path, Cost input_startup_cost, Cost input_total_cost, double tuples, int width)
{
Cost startup_cost = input_startup_cost;
Cost run_cost = input_total_cost - input_startup_cost;
int dop = SET_DOP(path->dop);
if (dop > 1) {
tuples = tuples / dop;
}
double nbytes = relation_byte_size(tuples, width, false, true, false);
long work_mem_bytes = u_sess->opt_cxt.op_work_mem * 1024L / dop;
/*
* Whether spilling or not, charge 2x cpu_operator_cost per tuple to
* reflect bookkeeping overhead. (This rate must be more than what
* cost_rescan charges for materialize, ie, cpu_operator_cost per tuple;
* if it is exactly the same then there will be a cost tie between
* nestloop with A outer, materialized B inner and nestloop with B outer,
* materialized A inner. The extra cost ensures we'll prefer
* materializing the smaller rel.) Note that this is normally a good deal
* less than cpu_tuple_cost; which is OK because a Material plan node
* doesn't do qual-checking or projection, so it's got less overhead than
* most plan nodes.
*/
run_cost += 2 * u_sess->attr.attr_sql.cpu_operator_cost * tuples;
/*
* If we will spill to disk, charge at the rate of seq_page_cost per page.
* This cost is assumed to be evenly spread through the plan run phase,
* which isn't exactly accurate but our cost model doesn't allow for
* nonuniform costs within the run phase.
*/
if (nbytes > work_mem_bytes) {
double npages = ceil(nbytes / BLCKSZ);
run_cost += u_sess->attr.attr_sql.seq_page_cost * npages;
}
path->startup_cost = startup_cost;
path->total_cost = startup_cost + run_cost;
}
/*
* Decide sonic hashagg routine or not.
* Similar with the same function in vsonichashagg.cpp
*/
bool isSonicHashAggPlanEnable(PlannerInfo* root, AggStrategy aggstrategy, int numGroupCols)
{
if (!(root->glob->vectorized) || aggstrategy != AGG_HASHED || !u_sess->attr.attr_sql.enable_sonic_hashagg)
return false;
/*
* Get the target list.
* Including all the corresponding columns in group by clause, whether or not it appears in target list.
*/
List* tleList = root->parse->targetList;
/* Get hash key in group by clause. */
List* sgClauses = root->parse->groupClause;
ListCell* lc = NULL;
Oid hashtype = 0;
foreach (lc, sgClauses) {
SortGroupClause* sortcl = (SortGroupClause*)lfirst(lc);
TargetEntry* tle = get_sortgroupclause_tle(sortcl, tleList, false);
if (tle == NULL)
return false;
switch (nodeTag(tle->expr)) {
case T_Var: {
Var* var = (Var*)(tle->expr);
hashtype = var->vartype;
} break;
case T_FuncExpr: {
FuncExpr* funcexpr = (FuncExpr*)(tle->expr);
hashtype = funcexpr->funcresulttype;
} break;
default:
break;
}
switch (hashtype) {
case CHAROID:
case BPCHAROID:
case INT1OID:
case INT2OID:
case INT4OID:
case INT8OID:
case FLOAT4OID:
case FLOAT8OID:
case TIMESTAMPOID:
case DATEOID:
case TEXTOID:
case VARCHAROID:
break;
default:
return false;
}
}
/*
* Aggregate function only support sum(), avg() function for int4, int8 and numeric tyep.
* Loop over all the targetlist and check aggref.
*/
foreach (lc, tleList) {
TargetEntry* tre = (TargetEntry*)lfirst(lc);
switch (nodeTag(tre->expr)) {
case T_Aggref: {
Aggref* aggref = (Aggref*)tre->expr;
if (!isAggrefSonicEnable(aggref->aggfnoid))
return false;
/* count(*) has no args */
if (aggref->aggfnoid == COUNTOID || aggref->aggfnoid == ANYCOUNTOID)
continue;
Expr* refexpr = (Expr*)linitial(aggref->args);
/* We only support simple expression cases */
if (!isExprSonicEnable(refexpr))
return false;
} break;
case T_FuncExpr: {
return false;
}
case T_Var:
case T_Const:
break;
default:
return false;
}
}
return true;
}
/*
* For operator HashAgg, compute appropriate size for hashtable given the estimated size of the
* columns to be hashed (number of rows).
*/
double estimate_hashagg_size(Path* path, PlannerInfo* root, AggStrategy aggstrategy, int numGroupCols, double numGroups,
double input_tuples, int input_width, int hash_entry_size)
{
double hash_table_size;
if (hash_entry_size == 0)
hash_entry_size = alloc_trunk_size(input_width) + HASH_ENTRY_OVERHEAD;
hash_table_size = (double)(hash_entry_size * numGroups) / 1024L;
return hash_table_size;
}
/*
* cost_agg
* Determines and returns the cost of performing an Agg plan node,
* including the cost of its input.
*
* aggcosts can be NULL when there are no actual aggregate functions (i.e.,
* we are using a hashed Agg node just to do grouping).
*
* Note: when aggstrategy == AGG_SORTED, caller must ensure that input costs
* are for appropriately-sorted input.
*/
void cost_agg(Path* path, PlannerInfo* root, AggStrategy aggstrategy, const AggClauseCosts* aggcosts, int numGroupCols,
double numGroups, Cost input_startup_cost, Cost input_total_cost, double input_tuples, int input_width,
int hash_entry_size, int dop, OpMemInfo* mem_info)
{
double output_tuples;
Cost startup_cost;
Cost total_cost;
AggClauseCosts dummy_aggcosts;
double hllagg_size = 0;
/* Use all-zero per-aggregate costs if NULL is passed */
if (aggcosts == NULL) {
AssertEreport(aggstrategy == AGG_HASHED,
MOD_OPT,
"Only support Hashed aggstrategy"
"when determining the cost of performing an Agg plan node.");
errno_t errorno = EOK;
errorno = memset_s(&dummy_aggcosts, sizeof(AggClauseCosts), 0, sizeof(AggClauseCosts));
securec_check(errorno, "\0", "\0");
aggcosts = &dummy_aggcosts;
}
#ifdef ENABLE_MULTIPLE_NODES
hllagg_size = estimate_hllagg_size(numGroups, root->parse->targetList);
#endif
dop = SET_DOP(dop);
/*
* The transCost.per_tuple component of aggcosts should be charged once
* per input tuple, corresponding to the costs of evaluating the aggregate
* transfns and their input expressions (with any startup cost of course
* charged but once). The finalCost component is charged once per output
* tuple, corresponding to the costs of evaluating the finalfns.
*
* If we are grouping, we charge an additional u_sess->attr.attr_sql.cpu_operator_cost per
* grouping column per input tuple for grouping comparisons.
*
* We will produce a single output tuple if not grouping, and a tuple per
* group otherwise. We charge u_sess->attr.attr_sql.cpu_tuple_cost for each output tuple.
*
* Note: in this cost model, AGG_SORTED and AGG_HASHED have exactly the
* same total CPU cost, but AGG_SORTED has lower startup cost. If the
* input path is already sorted appropriately, AGG_SORTED should be
* preferred (since it has no risk of memory overflow). This will happen
* as long as the computed total costs are indeed exactly equal --- but if
* there's roundoff error we might do the wrong thing. So be sure that
* the computations below form the same intermediate values in the same
* order.
*/
if (aggstrategy == AGG_PLAIN) {
startup_cost = input_total_cost;
startup_cost += aggcosts->transCost.startup;
startup_cost += aggcosts->transCost.per_tuple * input_tuples;
startup_cost += aggcosts->finalCost;
/* we aren't grouping */
total_cost = startup_cost + u_sess->attr.attr_sql.cpu_tuple_cost;
output_tuples = 1;
} else if (aggstrategy == AGG_SORTED) {
/* Here we are able to deliver output on-the-fly */
startup_cost = input_startup_cost;
total_cost = input_total_cost;
/* calcs phrased this way to match HASHED case, see note above */
total_cost += aggcosts->transCost.startup;
total_cost += aggcosts->transCost.per_tuple * input_tuples;
total_cost += (u_sess->attr.attr_sql.cpu_operator_cost * numGroupCols) * input_tuples;
total_cost += aggcosts->finalCost * numGroups;
total_cost += u_sess->attr.attr_sql.cpu_tuple_cost * numGroups;
output_tuples = numGroups;
} else {
bool spill_disk = false;
double hash_table_size;
/* must be AGG_HASHED */
startup_cost = input_total_cost;
startup_cost += aggcosts->transCost.startup;
startup_cost += aggcosts->transCost.per_tuple * input_tuples / dop;
startup_cost += (u_sess->attr.attr_sql.cpu_operator_cost * numGroupCols) * input_tuples / dop;
total_cost = startup_cost;
total_cost += aggcosts->finalCost * numGroups / dop;
total_cost += u_sess->attr.attr_sql.cpu_tuple_cost * numGroups / dop;
output_tuples = numGroups;
/* Get hash table size and estimate mem_info. */
hash_table_size = estimate_hashagg_size(
path, root, aggstrategy, numGroupCols, numGroups, input_tuples, input_width, hash_entry_size);
/* one more step estimate for hll */
hash_table_size += hllagg_size;
if (hash_table_size < 0)
hash_table_size = (double)LONG_MAX;
spill_disk = (hash_table_size > (double)u_sess->opt_cxt.op_work_mem);
if (spill_disk) {
const double disk_ratio = 1 - u_sess->opt_cxt.op_work_mem / hash_table_size;
double disk_pages = ceil(page_size(input_tuples, input_width) * disk_ratio);
double one_disk_io_cost = u_sess->attr.attr_sql.seq_page_cost * disk_pages;
double disk_hash_cost = u_sess->attr.attr_sql.cpu_operator_cost * numGroupCols * input_tuples * disk_ratio;
startup_cost += disk_hash_cost + one_disk_io_cost; /* hash, write cost counted */
total_cost += disk_hash_cost + 2 * one_disk_io_cost; /* hash, write, read cost counted */
}
if (mem_info != NULL) {
double disk_pages = ceil(page_size(input_tuples, input_width));
double one_disk_io_cost = u_sess->attr.attr_sql.seq_page_cost * disk_pages;
double disk_hash_cost = u_sess->attr.attr_sql.cpu_operator_cost * numGroupCols * input_tuples;
mem_info->opMem = u_sess->opt_cxt.op_work_mem;
mem_info->maxMem = hash_table_size;
mem_info->minMem = mem_info->maxMem / HASH_MAX_DISK_SIZE;
mem_info->regressCost = (disk_hash_cost + 2 * one_disk_io_cost);
}
}
path->rows = get_global_rows(output_tuples, 1.0, ng_get_dest_num_data_nodes(path));
path->multiple = 1.0;
path->startup_cost = startup_cost;
path->total_cost = total_cost;
}
/*
* cost_windowagg
* Determines and returns the cost of performing a WindowAgg plan node,
* including the cost of its input.
*
* Input is assumed already properly sorted.
*/
void cost_windowagg(Path* path, PlannerInfo* root, List* windowFuncs, int numPartCols, int numOrderCols,
Cost input_startup_cost, Cost input_total_cost, double input_tuples)
{
Cost startup_cost;
Cost total_cost;
ListCell* lc = NULL;
startup_cost = input_startup_cost;
total_cost = input_total_cost;
/*
* Window functions are assumed to cost their stated execution cost, plus
* the cost of evaluating their input expressions, per tuple. Since they
* may in fact evaluate their inputs at multiple rows during each cycle,
* this could be a drastic underestimate; but without a way to know how
* many rows the window function will fetch, it's hard to do better. In
* any case, it's a good estimate for all the built-in window functions,
* so we'll just do this for now.
*/
foreach (lc, windowFuncs) {
WindowFunc* wfunc = (WindowFunc*)lfirst(lc);
Cost wfunccost;
QualCost argcosts;
AssertEreport(IsA(wfunc, WindowFunc),
MOD_OPT,
"The nodeTag of wfunc is not T_WindowFunc"
"when determining the cost of performing a WindowAgg plan node.");
wfunccost = get_func_cost(wfunc->winfnoid) * u_sess->attr.attr_sql.cpu_operator_cost;
/* also add the input expressions' cost to per-input-row costs */
cost_qual_eval_node(&argcosts, (Node*)wfunc->args, root);
startup_cost += argcosts.startup;
wfunccost += argcosts.per_tuple;
total_cost += wfunccost * input_tuples;
}
/*
* We also charge cpu_operator_cost per grouping column per tuple for
* grouping comparisons, plus cpu_tuple_cost per tuple for general
* overhead.
*
* XXX this neglects costs of spooling the data to disk when it overflows
* work_mem. Sooner or later that should get accounted for.
*/
total_cost += u_sess->attr.attr_sql.cpu_operator_cost * (numPartCols + numOrderCols) * input_tuples;
total_cost += u_sess->attr.attr_sql.cpu_tuple_cost * input_tuples;
path->startup_cost = startup_cost;
path->total_cost = total_cost;
}
/*
* cost_group
* Determines and returns the cost of performing a Group plan node,
* including the cost of its input.
*
* Note: caller must ensure that input costs are for appropriately-sorted
* input.
*/
void cost_group(Path* path, PlannerInfo* root, int numGroupCols, double numGroups, Cost input_startup_cost,
Cost input_total_cost, double input_tuples)
{
Cost startup_cost;
Cost total_cost;
startup_cost = input_startup_cost;
total_cost = input_total_cost;
/*
* Charge one cpu_operator_cost per comparison per input tuple. We assume
* all columns get compared at most of the tuples.
*/
total_cost += u_sess->attr.attr_sql.cpu_operator_cost * input_tuples * numGroupCols;
path->rows = get_global_rows(numGroups, 1.0, ng_get_dest_num_data_nodes(path));
path->multiple = 1.0;
path->startup_cost = startup_cost;
path->total_cost = total_cost;
}
/*
* @Description: Calculate cost and adjust output rows for limit node.
*
* @param[IN] plan: limit plan
* @param[IN] lefttree: subplan
* @return void
*/
void cost_limit(Plan* plan, Plan* lefttree, int64 offset_est, int64 count_est)
{
/*
* Adjust the output rows count and costs according to the offset/limit.
* This is only a cosmetic issue if we are at top level, but if we are
* building a subquery then it's important to report correct info to the
* outer planner.
*
* When the offset or count couldn't be estimated, use 10% of the
* estimated number of rows emitted from the subplan.
*/
if (offset_est != 0) {
double offset_rows;
if (offset_est > 0) {
offset_rows = (double)offset_est;
if (is_replicated_plan(lefttree) && is_execute_on_datanodes(lefttree)) {
offset_rows *= ng_get_dest_num_data_nodes(lefttree);
}
} else
offset_rows = clamp_row_est(lefttree->plan_rows * 0.10);
if (offset_rows > lefttree->plan_rows)
offset_rows = lefttree->plan_rows;
if (plan->plan_rows > 0)
plan->startup_cost += (plan->total_cost - plan->startup_cost) * offset_rows / plan->plan_rows;
plan->plan_rows -= offset_rows;
if (plan->plan_rows < 1)
plan->plan_rows = 1;
}
if (count_est != 0) {
double count_rows;
if (count_est > 0) {
count_rows = (double)count_est;
if (is_execute_on_datanodes(lefttree)) {
count_rows *= ng_get_dest_num_data_nodes(lefttree);
}
} else
count_rows = clamp_row_est(lefttree->plan_rows * 0.10);
if (count_rows > plan->plan_rows)
count_rows = plan->plan_rows;
if (plan->plan_rows > 0)
plan->total_cost =
plan->startup_cost + (plan->total_cost - plan->startup_cost) * count_rows / plan->plan_rows;
plan->plan_rows = count_rows;
if (plan->plan_rows < 1)
plan->plan_rows = 1;
}
}
/*
* initial_cost_nestloop
* Preliminary estimate of the cost of a nestloop join path.
*
* This must quickly produce lower-bound estimates of the path's startup and
* total costs. If we are unable to eliminate the proposed path from
* consideration using the lower bounds, final_cost_nestloop will be called
* to obtain the final estimates.
*
* The exact division of labor between this function and final_cost_nestloop
* is private to them, and represents a tradeoff between speed of the initial
* estimate and getting a tight lower bound. We choose to not examine the
* join quals here, since that's by far the most expensive part of the
* calculations. The end result is that CPU-cost considerations must be
* left for the second phase.
*
* 'workspace' is to be filled with startup_cost, total_cost, and perhaps
* other data to be used by final_cost_nestloop
* 'jointype' is the type of join to be performed
* 'outer_path' is the outer input to the join
* 'inner_path' is the inner input to the join
* 'sjinfo' is extra info about the join for selectivity estimation
* 'semifactors' contains valid data if jointype is SEMI or ANTI
*/
void initial_cost_nestloop(PlannerInfo* root, JoinCostWorkspace* workspace, JoinType jointype, Path* outer_path,
Path* inner_path, SpecialJoinInfo* sjinfo, SemiAntiJoinFactors* semifactors, int dop)
{
Cost startup_cost = 0;
Cost run_cost = 0;
double outer_path_rows = PATH_LOCAL_ROWS(outer_path) / dop;
Cost inner_rescan_start_cost;
Cost inner_rescan_total_cost;
Cost inner_run_cost;
Cost inner_rescan_run_cost;
errno_t rc = 0;
rc = memset_s(&workspace->inner_mem_info, sizeof(OpMemInfo), 0, sizeof(OpMemInfo));
securec_check(rc, "\0", "\0");
/* estimate costs to rescan the inner relation */
cost_rescan(root, inner_path, &inner_rescan_start_cost, &inner_rescan_total_cost, &workspace->inner_mem_info);
/* cost of source data */
/*
* NOTE: clearly, we must pay both outer and inner paths' startup_cost
* before we can start returning tuples, so the join's startup cost is
* their sum. We'll also pay the inner path's rescan startup cost
* multiple times.
*/
startup_cost += outer_path->startup_cost + inner_path->startup_cost;
run_cost += outer_path->total_cost - outer_path->startup_cost;
if (outer_path_rows > 1)
run_cost += (outer_path_rows - 1) * inner_rescan_start_cost;
inner_run_cost = inner_path->total_cost - inner_path->startup_cost;
inner_rescan_run_cost = inner_rescan_total_cost - inner_rescan_start_cost;
if (jointype == JOIN_SEMI || jointype == JOIN_ANTI) {
double outer_matched_rows;
Selectivity inner_scan_frac;
/*
* SEMI or ANTI join: executor will stop after first match.
*
* For an outer-rel row that has at least one match, we can expect the
* inner scan to stop after a fraction 1/(match_count+1) of the inner
* rows, if the matches are evenly distributed. Since they probably
* aren't quite evenly distributed, we apply a fuzz factor of 2.0 to
* that fraction. (If we used a larger fuzz factor, we'd have to
* clamp inner_scan_frac to at most 1.0; but since match_count is at
* least 1, no such clamp is needed now.)
*
* A complicating factor is that rescans may be cheaper than first
* scans. If we never scan all the way to the end of the inner rel,
* it might be (depending on the plan type) that we'd never pay the
* whole inner first-scan run cost. However it is difficult to
* estimate whether that will happen, so be conservative and always
* charge the whole first-scan cost once.
*/
run_cost += inner_run_cost;
outer_matched_rows = rint(outer_path_rows * semifactors->outer_match_frac);
inner_scan_frac = 2.0 / (semifactors->match_count + 1.0);
/* Add inner run cost for additional outer tuples having matches */
if (outer_matched_rows > 1)
run_cost += (outer_matched_rows - 1) * inner_rescan_run_cost * inner_scan_frac;
/*
* The cost of processing unmatched rows varies depending on the
* details of the joinclauses, so we leave that part for later.
*/
/* Save private data for final_cost_nestloop */
workspace->outer_matched_rows = outer_matched_rows;
workspace->inner_scan_frac = inner_scan_frac;
workspace->inner_mem_info.regressCost *= Max(outer_matched_rows, 1.0);
} else {
/* Normal case; we'll scan whole input rel for each outer row */
run_cost += inner_run_cost;
if (outer_path_rows > 1)
run_cost += (outer_path_rows - 1) * inner_rescan_run_cost;
workspace->inner_mem_info.regressCost *= Max(outer_path_rows, 1.0);
}
/* CPU costs left for later */
/* Public result fields */
workspace->startup_cost = startup_cost;
workspace->total_cost = startup_cost + run_cost;
/* Save private data for final_cost_nestloop */
workspace->run_cost = run_cost;
workspace->inner_rescan_run_cost = inner_rescan_run_cost;
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("Initial nestloop cost: startup_cost: %lf, total_cost: %lf",
workspace->startup_cost,
workspace->total_cost)));
}
/*
* final_cost_nestloop
* Final estimate of the cost and result size of a nestloop join path.
*
* 'path' is already filled in except for the rows and cost fields
* 'workspace' is the result from initial_cost_nestloop
* 'sjinfo' is extra info about the join for selectivity estimation
* 'semifactors' contains valid data if path->jointype is SEMI or ANTI
*/
void final_cost_nestloop(PlannerInfo* root, NestPath* path, JoinCostWorkspace* workspace, SpecialJoinInfo* sjinfo,
SemiAntiJoinFactors* semifactors, bool hasalternative, int dop)
{
Path* outer_path = path->outerjoinpath;
Path* inner_path = path->innerjoinpath;
double outer_path_rows = PATH_LOCAL_ROWS(outer_path) / dop;
double inner_path_rows = PATH_LOCAL_ROWS(inner_path) / dop;
Cost startup_cost = workspace->startup_cost;
Cost run_cost = workspace->run_cost;
Cost inner_rescan_run_cost = workspace->inner_rescan_run_cost;
Cost cpu_per_tuple = 0.0;
QualCost restrict_qual_cost;
double ntuples;
bool ppi_used = false;
bool method_enabled = true;
/* Mark the path with the correct row estimate */
set_rel_path_rows(&path->path, path->path.parent, path->path.param_info);
/* For partial paths, scale row estimate. */
if (path->path.parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(&path->path);
path->path.rows = clamp_row_est(path->path.rows / parallel_divisor);
}
/*
* If inner_path or outer_path is EC functioinScan without stream,
* we should set the multiple particularly.
*/
set_joinpath_multiple_for_EC(root, &path->path, outer_path, inner_path);
/*
* We could include g_instance.cost_cxt.disable_cost in the preliminary estimate, but that
* would amount to optimizing for the case where the join method is
* disabled, which doesn't seem like the way to bet.
*/
if (inner_path->param_info != NULL && outer_path->parent != NULL)
ppi_used = bms_overlap(inner_path->param_info->ppi_req_outer, outer_path->parent->relids);
/* If ppi is used, we use enable_index_nestloop to judge whether use this path */
method_enabled = ppi_used ? u_sess->attr.attr_sql.enable_index_nestloop
: (u_sess->attr.attr_sql.enable_nestloop || !hasalternative);
if (!method_enabled)
startup_cost += g_instance.cost_cxt.disable_cost;
/* cost of source data */
if (path->jointype == JOIN_SEMI || path->jointype == JOIN_ANTI) {
double outer_matched_rows = workspace->outer_matched_rows;
Selectivity inner_scan_frac = workspace->inner_scan_frac;
/*
* SEMI or ANTI join: executor will stop after first match.
*/
/* Compute number of tuples processed (not number emitted!) */
ntuples = outer_matched_rows * inner_path_rows * inner_scan_frac;
/*
* For unmatched outer-rel rows, there are two cases. If the inner
* path is an indexscan using all the joinquals as indexquals, then an
* unmatched row results in an indexscan returning no rows, which is
* probably quite cheap. We estimate this case as the same cost to
* return the first tuple of a nonempty scan. Otherwise, the executor
* will have to scan the whole inner rel; not so cheap.
*/
if (has_indexed_join_quals(path)) {
run_cost += (outer_path_rows - outer_matched_rows) * inner_rescan_run_cost / inner_path_rows;
/*
* We won't be evaluating any quals at all for these rows, so
* don't add them to ntuples.
*/
} else {
run_cost += (outer_path_rows - outer_matched_rows) * inner_rescan_run_cost;
ntuples += (outer_path_rows - outer_matched_rows) * inner_path_rows;
}
} else {
/* Normal-case source costs were included in preliminary estimate */
/* Compute number of tuples processed (not number emitted!) */
ntuples = outer_path_rows * inner_path_rows;
}
/* CPU costs */
cost_qual_eval(&restrict_qual_cost, path->joinrestrictinfo, root);
startup_cost += restrict_qual_cost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + restrict_qual_cost.per_tuple;
run_cost += cpu_per_tuple * ntuples;
path->path.startup_cost = startup_cost;
path->path.total_cost = startup_cost + run_cost;
path->path.stream_cost = outer_path->stream_cost;
if (!method_enabled)
path->path.total_cost *= g_instance.cost_cxt.disable_cost_enlarge_factor;
if (path->innerjoinpath->pathtype == T_Material)
copy_mem_info(&((MaterialPath*)path->innerjoinpath)->mem_info, &workspace->inner_mem_info);
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("final cost nest loop: stream_cost: %lf, startup_cost: %lf, total_cost: %lf",
path->path.stream_cost,
path->path.startup_cost,
path->path.total_cost)));
}
/*
* initial_cost_mergejoin
* Preliminary estimate of the cost of a mergejoin path.
*
* This must quickly produce lower-bound estimates of the path's startup and
* total costs. If we are unable to eliminate the proposed path from
* consideration using the lower bounds, final_cost_mergejoin will be called
* to obtain the final estimates.
*
* The exact division of labor between this function and final_cost_mergejoin
* is private to them, and represents a tradeoff between speed of the initial
* estimate and getting a tight lower bound. We choose to not examine the
* join quals here, except for obtaining the scan selectivity estimate which
* is really essential (but fortunately, use of caching keeps the cost of
* getting that down to something reasonable).
* We also assume that cost_sort is cheap enough to use here.
*
* 'workspace' is to be filled with startup_cost, total_cost, and perhaps
* other data to be used by final_cost_mergejoin
* 'jointype' is the type of join to be performed
* 'mergeclauses' is the list of joinclauses to be used as merge clauses
* 'outer_path' is the outer input to the join
* 'inner_path' is the inner input to the join
* 'outersortkeys' is the list of sort keys for the outer path
* 'innersortkeys' is the list of sort keys for the inner path
* 'sjinfo' is extra info about the join for selectivity estimation
*
* Note: outersortkeys and innersortkeys should be NIL if no explicit
* sort is needed because the respective source path is already ordered.
*/
void initial_cost_mergejoin(PlannerInfo* root, JoinCostWorkspace* workspace, JoinType jointype, List* mergeclauses,
Path* outer_path, Path* inner_path, List* outersortkeys, List* innersortkeys, SpecialJoinInfo* sjinfo)
{
Cost startup_cost = 0;
Cost run_cost = 0;
double outer_path_rows = PATH_LOCAL_ROWS(outer_path);
double inner_path_rows = PATH_LOCAL_ROWS(inner_path);
Cost inner_run_cost;
double outer_rows, inner_rows, outer_skip_rows, inner_skip_rows;
Selectivity outerstartsel, outerendsel, innerstartsel, innerendsel;
Path sort_path; /* dummy for result of cost_sort */
errno_t rc = 0;
rc = memset_s(&workspace->outer_mem_info, sizeof(OpMemInfo), 0, sizeof(OpMemInfo));
securec_check(rc, "\0", "\0");
rc = memset_s(&workspace->inner_mem_info, sizeof(OpMemInfo), 0, sizeof(OpMemInfo));
securec_check(rc, "\0", "\0");
/* Protect some assumptions below that rowcounts aren't zero or NaN */
if (outer_path_rows <= 0 || isnan(outer_path_rows))
outer_path_rows = 1;
if (inner_path_rows <= 0 || isnan(inner_path_rows))
inner_path_rows = 1;
/*
* A merge join will stop as soon as it exhausts either input stream
* (unless it's an outer join, in which case the outer side has to be
* scanned all the way anyway). Estimate fraction of the left and right
* inputs that will actually need to be scanned. Likewise, we can
* estimate the number of rows that will be skipped before the first join
* pair is found, which should be factored into startup cost. We use only
* the first (most significant) merge clause for this purpose. Since
* mergejoinscansel() is a fairly expensive computation, we cache the
* results in the merge clause RestrictInfo.
*/
if (mergeclauses != NIL && jointype != JOIN_FULL) {
RestrictInfo* firstclause = (RestrictInfo*)linitial(mergeclauses);
List* opathkeys = NIL;
List* ipathkeys = NIL;
PathKey* opathkey = NULL;
PathKey* ipathkey = NULL;
MergeScanSelCache* cache = NULL;
/* Get the input pathkeys to determine the sort-order details */
opathkeys = outersortkeys ? outersortkeys : outer_path->pathkeys;
ipathkeys = innersortkeys ? innersortkeys : inner_path->pathkeys;
AssertEreport(opathkeys != NIL,
MOD_OPT,
"The outer pathkeys is null when determining the cost of performing a mergejoin path.");
AssertEreport(ipathkeys != NIL,
MOD_OPT,
"The inner pathkeys is null when determining the cost of performing a mergejoin path.");
opathkey = (PathKey*)linitial(opathkeys);
ipathkey = (PathKey*)linitial(ipathkeys);
/* debugging check */
if (opathkey->pk_opfamily != ipathkey->pk_opfamily ||
opathkey->pk_eclass->ec_collation != ipathkey->pk_eclass->ec_collation ||
opathkey->pk_strategy != ipathkey->pk_strategy || opathkey->pk_nulls_first != ipathkey->pk_nulls_first)
ereport(ERROR,
(errmodule(MOD_OPT),
errcode(ERRCODE_OPTIMIZER_INCONSISTENT_STATE),
errmsg("left and right pathkeys do not match in mergejoin when initlize cost")));
/* Get the selectivity with caching */
cache = cached_scansel(root, firstclause, opathkey);
if (bms_is_subset(firstclause->left_relids, outer_path->parent->relids)) {
/* left side of clause is outer */
outerstartsel = cache->leftstartsel;
outerendsel = cache->leftendsel;
innerstartsel = cache->rightstartsel;
innerendsel = cache->rightendsel;
} else {
/* left side of clause is inner */
outerstartsel = cache->rightstartsel;
outerendsel = cache->rightendsel;
innerstartsel = cache->leftstartsel;
innerendsel = cache->leftendsel;
}
if (jointype == JOIN_LEFT || jointype == JOIN_LEFT_ANTI_FULL || jointype == JOIN_ANTI) {
outerstartsel = 0.0;
outerendsel = 1.0;
} else if (jointype == JOIN_RIGHT) {
innerstartsel = 0.0;
innerendsel = 1.0;
}
/* jointype should not be JOIN_RIGHT_ANTI_FULL,
* because JOIN_RIGHT_ANTI_FULL can not create a mergejoin plan.
*/
AssertEreport(jointype != JOIN_RIGHT_ANTI_FULL,
MOD_OPT,
"The mergejoin plan with JOIN_RIGHT_ANTI_FULL is not allowed."
"when determining the cost of performing a mergejoin path.");
} else {
/* cope with clauseless or full mergejoin */
outerstartsel = innerstartsel = 0.0;
outerendsel = innerendsel = 1.0;
}
/*
* Convert selectivities to row counts. We force outer_rows and
* inner_rows to be at least 1, but the skip_rows estimates can be zero.
*/
outer_skip_rows = rint(outer_path_rows * outerstartsel);
inner_skip_rows = rint(inner_path_rows * innerstartsel);
outer_rows = clamp_row_est(outer_path_rows * outerendsel);
inner_rows = clamp_row_est(inner_path_rows * innerendsel);
AssertEreport(outer_skip_rows <= outer_rows,
MOD_OPT,
"The estimated skip rows is larger than rounding outer rows which avoid possible divide-by-zero "
"when determining the cost of performing a mergejoin path.");
AssertEreport(inner_skip_rows <= inner_rows,
MOD_OPT,
"The estimated skip rows is larger than rounding inner rows which avoid possible divide-by-zero"
"when determining the cost of performing a mergejoin path.");
/*
* Readjust scan selectivities to account for above rounding. This is
* normally an insignificant effect, but when there are only a few rows in
* the inputs, failing to do this makes for a large percentage error.
*/
outerstartsel = outer_skip_rows / outer_path_rows;
innerstartsel = inner_skip_rows / inner_path_rows;
outerendsel = outer_rows / outer_path_rows;
innerendsel = inner_rows / inner_path_rows;
AssertEreport(outerstartsel <= outerendsel,
MOD_OPT,
"The selectivities corresponding to estimated skip rows is larger than that of above rounding outer rows"
"when determining the cost of performing a mergejoin path.");
AssertEreport(innerstartsel <= innerendsel,
MOD_OPT,
"The selectivities corresponding to estimated skip rows is larger than that of above rounding inner rows"
"when determining the cost of performing a mergejoin path.");
/* cost of source data */
if (outersortkeys || IsA(outer_path, StreamPath)) { /* do we need to sort outer? */
int outer_width = get_path_actual_total_width(outer_path, root->glob->vectorized, OP_SORT);
cost_sort(&sort_path,
outersortkeys,
outer_path->total_cost,
outer_path_rows,
outer_width,
0.0,
u_sess->opt_cxt.op_work_mem,
-1.0,
root->glob->vectorized,
1,
&workspace->outer_mem_info);
startup_cost += sort_path.startup_cost;
startup_cost += (sort_path.total_cost - sort_path.startup_cost) * outerstartsel;
run_cost += (sort_path.total_cost - sort_path.startup_cost) * (outerendsel - outerstartsel);
} else {
startup_cost += outer_path->startup_cost;
startup_cost += (outer_path->total_cost - outer_path->startup_cost) * outerstartsel;
run_cost += (outer_path->total_cost - outer_path->startup_cost) * (outerendsel - outerstartsel);
}
if (innersortkeys || IsA(inner_path, StreamPath)) { /* do we need to sort inner? */
int inner_width = get_path_actual_total_width(inner_path, root->glob->vectorized, OP_SORT);
cost_sort(&sort_path,
innersortkeys,
inner_path->total_cost,
inner_path_rows,
inner_width,
0.0,
u_sess->opt_cxt.op_work_mem,
-1.0,
root->glob->vectorized,
1,
&workspace->inner_mem_info);
startup_cost += sort_path.startup_cost;
startup_cost += (sort_path.total_cost - sort_path.startup_cost) * innerstartsel;
inner_run_cost = (sort_path.total_cost - sort_path.startup_cost) * (innerendsel - innerstartsel);
} else {
startup_cost += inner_path->startup_cost;
startup_cost += (inner_path->total_cost - inner_path->startup_cost) * innerstartsel;
inner_run_cost = (inner_path->total_cost - inner_path->startup_cost) * (innerendsel - innerstartsel);
}
/*
* We can't yet determine whether rescanning occurs, or whether
* materialization of the inner input should be done. The minimum
* possible inner input cost, regardless of rescan and materialization
* considerations, is inner_run_cost. We include that in
* workspace->total_cost, but not yet in run_cost.
*/
/* CPU costs left for later */
/* Public result fields */
workspace->startup_cost = startup_cost;
workspace->total_cost = startup_cost + run_cost + inner_run_cost;
/* Save private data for final_cost_mergejoin */
workspace->run_cost = run_cost;
workspace->inner_run_cost = inner_run_cost;
workspace->outer_rows = outer_rows;
workspace->inner_rows = inner_rows;
workspace->outer_skip_rows = outer_skip_rows;
workspace->inner_skip_rows = inner_skip_rows;
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("Initial mergejoin cost: startup_cost: %lf, total_cost: %lf",
workspace->startup_cost,
workspace->total_cost)));
}
/*
* final_cost_mergejoin
* Final estimate of the cost and result size of a mergejoin path.
*
* Unlike other costsize functions, this routine makes one actual decision:
* whether we should materialize the inner path. We do that either because
* the inner path can't support mark/restore, or because it's cheaper to
* use an interposed Material node to handle mark/restore. When the decision
* is cost-based it would be logically cleaner to build and cost two separate
* paths with and without that flag set; but that would require repeating most
* of the cost calculations, which are not all that cheap. Since the choice
* will not affect output pathkeys or startup cost, only total cost, there is
* no possibility of wanting to keep both paths. So it seems best to make
* the decision here and record it in the path's materialize_inner field.
*
* 'path' is already filled in except for the rows and cost fields and
* materialize_inner
* 'workspace' is the result from initial_cost_mergejoin
* 'sjinfo' is extra info about the join for selectivity estimation
*/
void final_cost_mergejoin(
PlannerInfo* root, MergePath* path, JoinCostWorkspace* workspace, SpecialJoinInfo* sjinfo, bool hasalternative)
{
Path* outer_path = path->jpath.outerjoinpath;
Path* inner_path = path->jpath.innerjoinpath;
double inner_path_rows = PATH_LOCAL_ROWS(inner_path);
List* mergeclauses = path->path_mergeclauses;
List* innersortkeys = path->innersortkeys;
Cost startup_cost = workspace->startup_cost;
Cost run_cost = workspace->run_cost;
Cost inner_run_cost = workspace->inner_run_cost;
double outer_rows = workspace->outer_rows;
double inner_rows = workspace->inner_rows;
double outer_skip_rows = workspace->outer_skip_rows;
double inner_skip_rows = workspace->inner_skip_rows;
Cost cpu_per_tuple, bare_inner_cost, mat_inner_cost;
QualCost merge_qual_cost;
QualCost qp_qual_cost;
double mergejointuples, rescannedtuples;
double rescanratio;
/* Protect some assumptions below that rowcounts aren't zero or NaN */
if (inner_path_rows <= 0 || isnan(inner_path_rows))
inner_path_rows = 1;
/* Mark the path with the correct row estimate */
set_rel_path_rows(&path->jpath.path, path->jpath.path.parent, path->jpath.path.param_info);
/* For partial paths, scale row estimate. */
if (path->jpath.path.parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(&path->jpath.path);
path->jpath.path.rows = clamp_row_est(path->jpath.path.rows / parallel_divisor);
}
/*
* If inner_path or outer_path is EC functioinScan without stream,
* we should set the multiple particularly.
*/
set_joinpath_multiple_for_EC(root, &path->jpath.path, outer_path, inner_path);
/*
* We could include g_instance.cost_cxt.disable_cost in the preliminary estimate, but that
* would amount to optimizing for the case where the join method is
* disabled, which doesn't seem like the way to bet.
*/
if (!u_sess->attr.attr_sql.enable_mergejoin && hasalternative)
startup_cost += g_instance.cost_cxt.disable_cost;
/*
* Compute cost of the mergequals and qpquals (other restriction clauses)
* separately.
*/
cost_qual_eval(&merge_qual_cost, mergeclauses, root);
cost_qual_eval(&qp_qual_cost, path->jpath.joinrestrictinfo, root);
qp_qual_cost.startup -= merge_qual_cost.startup;
qp_qual_cost.per_tuple -= merge_qual_cost.per_tuple;
/*
* Get approx # tuples passing the mergequals. We use approx_tuple_count
* here because we need an estimate done with JOIN_INNER semantics.
*/
mergejointuples = approx_tuple_count(root, &path->jpath, mergeclauses);
/*
* When there are equal merge keys in the outer relation, the mergejoin
* must rescan any matching tuples in the inner relation. This means
* re-fetching inner tuples; we have to estimate how often that happens.
*
* For regular inner and outer joins, the number of re-fetches can be
* estimated approximately as size of merge join output minus size of
* inner relation. Assume that the distinct key values are 1, 2, ..., and
* denote the number of values of each key in the outer relation as m1,
* m2, ...; in the inner relation, n1, n2, ... Then we have
*
* size of join = m1 * n1 + m2 * n2 + ...
*
* number of rescanned tuples = (m1 - 1) * n1 + (m2 - 1) * n2 + ... = m1 *
* n1 + m2 * n2 + ... - (n1 + n2 + ...) = size of join - size of inner
* relation
*
* This equation works correctly for outer tuples having no inner match
* (nk = 0), but not for inner tuples having no outer match (mk = 0); we
* are effectively subtracting those from the number of rescanned tuples,
* when we should not. Can we do better without expensive selectivity
* computations?
*
* The whole issue is moot if we are working from a unique-ified outer
* input.
*/
if (IsA(outer_path, UniquePath))
rescannedtuples = 0;
else {
rescannedtuples = mergejointuples - inner_path_rows;
/* Must clamp because of possible underestimate */
if (rescannedtuples < 0) {
rescannedtuples = 0;
}
}
/* We'll inflate various costs this much to account for rescanning */
rescanratio = 1.0 + (rescannedtuples / inner_path_rows);
/*
* Decide whether we want to materialize the inner input to shield it from
* mark/restore and performing re-fetches. Our cost model for regular
* re-fetches is that a re-fetch costs the same as an original fetch,
* which is probably an overestimate; but on the other hand we ignore the
* bookkeeping costs of mark/restore. Not clear if it's worth developing
* a more refined model. So we just need to inflate the inner run cost by
* rescanratio.
*/
bare_inner_cost = inner_run_cost * rescanratio;
/*
* When we interpose a Material node the re-fetch cost is assumed to be
* just cpu_operator_cost per tuple, independently of the underlying
* plan's cost; and we charge an extra cpu_operator_cost per original
* fetch as well. Note that we're assuming the materialize node will
* never spill to disk, since it only has to remember tuples back to the
* last mark. (If there are a huge number of duplicates, our other cost
* factors will make the path so expensive that it probably won't get
* chosen anyway.) So we don't use cost_rescan here.
*
* Note: keep this estimate in sync with create_mergejoin_plan's labeling
* of the generated Material node.
*/
mat_inner_cost = inner_run_cost + u_sess->attr.attr_sql.cpu_operator_cost * inner_path_rows * rescanratio;
int inner_width = get_path_actual_total_width(inner_path, root->glob->vectorized, OP_MATERIAL);
double inner_rel_size = relation_byte_size(inner_path_rows, inner_width, root->glob->vectorized, true, false);
bool mat_spill_disk = (inner_rel_size > (u_sess->opt_cxt.op_work_mem * 1024L));
/*
* Prefer materializing if it looks cheaper, unless the user has asked to
* suppress materialization.
*/
if (u_sess->attr.attr_sql.enable_material && mat_inner_cost < bare_inner_cost)
path->materialize_inner = true;
/*
* Even if materializing doesn't look cheaper, we *must* do it if the
* inner path is to be used directly (without sorting) and it doesn't
* support mark/restore.
*
* Since the inner side must be ordered, and only Sorts and IndexScans can
* create order to begin with, and they both support mark/restore, you
* might think there's no problem --- but you'd be wrong. Nestloop and
* merge joins can *preserve* the order of their inputs, so they can be
* selected as the input of a mergejoin, and they don't support
* mark/restore at present.
*
* We don't test the value of enable_material here, because
* materialization is required for correctness in this case, and turning
* it off does not entitle us to deliver an invalid plan.
*/
else if (innersortkeys == NIL && !ExecSupportsMarkRestore(inner_path->pathtype))
path->materialize_inner = true;
/*
* Also, force materializing if the inner path is to be sorted and the
* sort is expected to spill to disk. This is because the final merge
* pass can be done on-the-fly if it doesn't have to support mark/restore.
* We don't try to adjust the cost estimates for this consideration,
* though.
*
* Since materialization is a performance optimization in this case,
* rather than necessary for correctness, we skip it if enable_material is
* off.
*/
else if (u_sess->attr.attr_sql.enable_material && innersortkeys != NIL && mat_spill_disk)
path->materialize_inner = true;
else
path->materialize_inner = false;
/* Charge the right incremental cost for the chosen case */
if (path->materialize_inner) {
path->mat_mem_info.opMem = u_sess->opt_cxt.op_work_mem;
path->mat_mem_info.maxMem = inner_rel_size / 1024L;
path->mat_mem_info.minMem = path->mat_mem_info.maxMem / SORT_MAX_DISK_SIZE;
path->mat_mem_info.regressCost =
ceil(inner_rel_size / BLCKSZ) / SET_DOP(path->jpath.path.dop) * u_sess->attr.attr_sql.seq_page_cost * 2.0;
run_cost += mat_inner_cost;
} else
run_cost += bare_inner_cost;
/* CPU costs */
/*
* The number of tuple comparisons needed is approximately number of outer
* rows plus number of inner rows plus number of rescanned tuples (can we
* refine this?). At each one, we need to evaluate the mergejoin quals.
*/
startup_cost += merge_qual_cost.startup;
startup_cost += merge_qual_cost.per_tuple * (outer_skip_rows + inner_skip_rows * rescanratio);
run_cost +=
merge_qual_cost.per_tuple * ((outer_rows - outer_skip_rows) + (inner_rows - inner_skip_rows) * rescanratio);
/*
* For each tuple that gets through the mergejoin proper, we charge
* cpu_tuple_cost plus the cost of evaluating additional restriction
* clauses that are to be applied at the join. (This is pessimistic since
* not all of the quals may get evaluated at each tuple.)
*
* Note: we could adjust for SEMI/ANTI joins skipping some qual
* evaluations here, but it's probably not worth the trouble.
*/
startup_cost += qp_qual_cost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qp_qual_cost.per_tuple;
run_cost += cpu_per_tuple * mergejointuples;
copy_mem_info(&path->outer_mem_info, &workspace->outer_mem_info);
copy_mem_info(&path->inner_mem_info, &workspace->inner_mem_info);
path->jpath.path.startup_cost = startup_cost;
path->jpath.path.total_cost = startup_cost + run_cost;
path->jpath.path.stream_cost = outer_path->stream_cost;
if (!u_sess->attr.attr_sql.enable_mergejoin && hasalternative)
path->jpath.path.total_cost *= g_instance.cost_cxt.disable_cost_enlarge_factor;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("final cost merge join: stream_cost: %lf, startup_cost: %lf, total_cost: %lf",
path->jpath.path.stream_cost,
path->jpath.path.startup_cost,
path->jpath.path.total_cost)));
}
/*
* run mergejoinscansel() with caching
*/
MergeScanSelCache* cached_scansel(PlannerInfo* root, RestrictInfo* rinfo, PathKey* pathkey)
{
MergeScanSelCache* cache = NULL;
ListCell* lc = NULL;
Selectivity leftstartsel, leftendsel, rightstartsel, rightendsel;
MemoryContext oldcontext;
/* Do we have this result already? */
foreach (lc, rinfo->scansel_cache) {
cache = (MergeScanSelCache*)lfirst(lc);
if (cache->opfamily == pathkey->pk_opfamily && cache->collation == pathkey->pk_eclass->ec_collation &&
cache->strategy == pathkey->pk_strategy && cache->nulls_first == pathkey->pk_nulls_first)
return cache;
}
/* Nope, do the computation */
mergejoinscansel(root,
(Node*)rinfo->clause,
pathkey->pk_opfamily,
pathkey->pk_strategy,
pathkey->pk_nulls_first,
&leftstartsel,
&leftendsel,
&rightstartsel,
&rightendsel);
/* Cache the result in suitably long-lived workspace */
oldcontext = MemoryContextSwitchTo(root->planner_cxt);
cache = (MergeScanSelCache*)palloc(sizeof(MergeScanSelCache));
cache->opfamily = pathkey->pk_opfamily;
cache->collation = pathkey->pk_eclass->ec_collation;
cache->strategy = pathkey->pk_strategy;
cache->nulls_first = pathkey->pk_nulls_first;
cache->leftstartsel = leftstartsel;
cache->leftendsel = leftendsel;
cache->rightstartsel = rightstartsel;
cache->rightendsel = rightendsel;
rinfo->scansel_cache = lappend(rinfo->scansel_cache, cache);
MemoryContextSwitchTo(oldcontext);
return cache;
}
/*
* initial_cost_hashjoin
* Preliminary estimate of the cost of a hashjoin path.
*
* This must quickly produce lower-bound estimates of the path's startup and
* total costs. If we are unable to eliminate the proposed path from
* consideration using the lower bounds, final_cost_hashjoin will be called
* to obtain the final estimates.
*
* The exact division of labor between this function and final_cost_hashjoin
* is private to them, and represents a tradeoff between speed of the initial
* estimate and getting a tight lower bound. We choose to not examine the
* join quals here (other than by counting the number of hash clauses),
* so we can't do much with CPU costs. We do assume that
* ExecChooseHashTableSize is cheap enough to use here.
*
* 'workspace' is to be filled with startup_cost, total_cost, and perhaps
* other data to be used by final_cost_hashjoin
* 'jointype' is the type of join to be performed
* 'hashclauses' is the list of joinclauses to be used as hash clauses
* 'outer_path' is the outer input to the join
* 'inner_path' is the inner input to the join
* 'sjinfo' is extra info about the join for selectivity estimation
* 'semifactors' contains valid data if jointype is SEMI or ANTI
*/
void initial_cost_hashjoin(PlannerInfo* root, JoinCostWorkspace* workspace, JoinType jointype, List* hashclauses,
Path* outer_path, Path* inner_path, SpecialJoinInfo* sjinfo, SemiAntiJoinFactors* semifactors, int dop,
bool parallel_hash)
{
Cost startup_cost = 0;
Cost run_cost = 0;
double outer_path_rows = PATH_LOCAL_ROWS(outer_path) / dop;
double inner_path_rows = PATH_LOCAL_ROWS(inner_path) / dop;
double inner_path_rows_total = inner_path_rows;
int num_hashclauses = list_length(hashclauses);
int numbuckets;
int numbatches;
int num_skew_mcvs;
int inner_width; /* width of inner rel */
int outer_width; /* width of outer rel */
double outerpages;
double innerpages;
size_t space_allowed; /* unused */
errno_t rc = 0;
rc = memset_s(&workspace->inner_mem_info, sizeof(OpMemInfo), 0, sizeof(OpMemInfo));
securec_check(rc, "\0", "\0");
ereport(
DEBUG1, (errmodule(MOD_OPT_JOIN), errmsg("outer: %lf, %lf", outer_path->startup_cost, outer_path->total_cost)));
ereport(
DEBUG1, (errmodule(MOD_OPT_JOIN), errmsg("inner: %lf, %lf", inner_path->startup_cost, inner_path->total_cost)));
/* cost of source data */
startup_cost += outer_path->startup_cost;
run_cost += outer_path->total_cost - outer_path->startup_cost;
/*
* If this is a parallel hash build, then the value we have for
* inner_rows_total currently refers only to the rows returned by each
* participant. For shared hash table size estimation, we need the total
* number, so we need to undo the division.
*/
if (u_sess->attr.attr_sql.enable_parallel_hash) {
inner_path_rows_total *= get_parallel_divisor(inner_path);
}
/*
* Sometimes, we suffers the case that small table with large cost join
* with a large table. In such case, the cost mainly comes from large cost
* of small table, and join cost become similar (within 1% gap). Since 1%
* gap is tolerated during add_path(), we should temporarily remove large
* cost for small table to ensure the small table be the inner side for better
* performance, and then restore the cost back before do the final decision.
*/
if (!u_sess->attr.attr_sql.enable_change_hjcost)
startup_cost += inner_path->total_cost;
else {
startup_cost += inner_path->startup_cost;
run_cost += inner_path->total_cost - inner_path->startup_cost;
}
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Source data cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
/* save startup cost for later stream cut off */
Cost startup_cost_origin = startup_cost;
/* sours
* Cost of computing hash function: must do it once per input tuple. We
* charge one cpu_operator_cost for each column's hash function. Also,
* tack on one cpu_tuple_cost per inner row, to model the costs of
* inserting the row into the hashtable.
*
* XXX when a hashclause is more complex than a single operator, we really
* should charge the extra eval costs of the left or right side, as
* appropriate, here. This seems more work than it's worth at the moment.
*/
startup_cost += (u_sess->attr.attr_sql.cpu_operator_cost * num_hashclauses + u_sess->attr.attr_sql.cpu_tuple_cost +
u_sess->attr.attr_sql.allocate_mem_cost) *
inner_path_rows;
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * num_hashclauses * outer_path_rows;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add hash function cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
bool isComplicateHashKey = has_complicate_hashkey(hashclauses, inner_path->parent->relids);
int newcolnum = isComplicateHashKey ? 1 : 0;
/* for vectorized right join or right anti join, we should add more column to flag match or not */
if (root->glob->vectorized && (jointype == JOIN_RIGHT || jointype == JOIN_RIGHT_ANTI ||
jointype == JOIN_RIGHT_SEMI || jointype == JOIN_RIGHT_ANTI_FULL))
newcolnum++;
/*
* Get hash table size that executor would use for inner relation.
*
* XXX for the moment, always assume that skew optimization will be
* performed. As long as SKEW_WORK_MEM_PERCENT is small, it's not worth
* trying to determine that for sure.
*
* XXX at some point it might be interesting to try to account for skew
* optimization in the cost estimate, but for now, we don't.
*/
inner_width = get_path_actual_total_width(inner_path, root->glob->vectorized, OP_HASHJOIN, newcolnum);
outer_width = get_path_actual_total_width(outer_path, root->glob->vectorized, OP_HASHJOIN);
ExecChooseHashTableSize(inner_path_rows_total,
inner_width,
true,
parallel_hash, /* try_combined_work_mem */
outer_path->parallel_workers,
&space_allowed,
&numbuckets,
&numbatches,
&num_skew_mcvs,
u_sess->opt_cxt.op_work_mem / dop,
root->glob->vectorized,
&workspace->inner_mem_info);
innerpages = page_size(PATH_LOCAL_ROWS(inner_path), inner_width) / dop;
outerpages = page_size(PATH_LOCAL_ROWS(outer_path), outer_width) / dop;
/*
* If inner relation is too big then we will need to "batch" the join,
* which implies writing and reading most of the tuples to disk an extra
* time. Charge seq_page_cost per page, since the I/O should be nice and
* sequential. Writing the inner rel counts as startup cost, all the rest
* as run cost.
*/
double startuppagecost = u_sess->attr.attr_sql.seq_page_cost * innerpages;
double runpagecost = u_sess->attr.attr_sql.seq_page_cost * (innerpages + 2 * outerpages);
if (numbatches > 1) {
startup_cost += startuppagecost;
run_cost += runpagecost;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add seq page cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
}
/* cut stream_cost */
if (numbatches <= 1 /* don't cut stream cost if inner rel spill to disk */
&& outer_path->stream_cost > STREAM_COST_THRESHOLD && outer_path->stream_cost > 0.25 * inner_path->total_cost) {
Cost cut_cost = 0.75 * Min(outer_path->stream_cost, inner_path->total_cost);
if (cut_cost <= startup_cost_origin)
startup_cost -= cut_cost;
else {
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Abnormal case: outer_stream_cost: %lf > current startup_cost: %lf",
outer_path->stream_cost,
startup_cost)));
}
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("outer_stream_cost: %lf, after cut-off: startup_cost: %lf, run_cost: %lf, cut_off_cost: %lf",
outer_path->stream_cost,
startup_cost,
run_cost,
cut_cost)));
}
/* Set mem info for hash join path */
workspace->inner_mem_info.maxMem *= dop;
workspace->inner_mem_info.minMem = workspace->inner_mem_info.maxMem / HASH_MAX_DISK_SIZE;
workspace->inner_mem_info.opMem = u_sess->opt_cxt.op_work_mem;
workspace->inner_mem_info.regressCost = (startuppagecost + runpagecost);
/* CPU costs left for later */
/* Public result fields */
workspace->startup_cost = startup_cost;
workspace->total_cost = startup_cost + run_cost;
/* Save private data for final_cost_hashjoin */
workspace->run_cost = run_cost;
workspace->numbuckets = numbuckets;
workspace->numbatches = numbatches;
workspace->inner_rows_total = inner_path_rows_total;
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("Initial hashjoin cost: startup_cost: %lf, total_cost: %lf, work_mem: %d, esti_work_mem: %d",
workspace->startup_cost,
workspace->total_cost,
u_sess->opt_cxt.op_work_mem,
root->glob->estiopmem)));
}
/*
* compute_bucket_size
*
* acquire hashjoin comparison selectivity within one bucket. Here are some key points:
* 1. Since the stat should be acquired from system catalog, we should cache the final selectivity to
* decrease the times to fetch system cache, as pg does.
* 2. In distribute environment, selectivity differs a lot for non-stream, redistribute and broadcast
* case, so we should cache them respectively.
* 3. Since the minimum selectivity is limited by virtualbucket size, so we should also cache virtual
* bucket size, and recalculate the selectivity when virtualbucket size changes.
*
* sjinfo: identify join info include lefthand/righthand in order to judge if can use possion to estimate distinct.
*/
Selectivity compute_bucket_size(PlannerInfo* root, RestrictInfo* restrictinfo, double virtualbuckets, Path* inner_path,
bool left, SpecialJoinInfo* sjinfo, double* ndistinct)
{
Path* inner = NULL;
Selectivity thisbucketsize = -1;
BucketSize* bucket = left ? &restrictinfo->left_bucketsize : &restrictinfo->right_bucketsize;
if (!IsA(inner_path, StreamPath)) {
/* for replicate path, we should also adjust to global distinct value */
if (IsA(inner_path, HashPath) || IsLocatorReplicated(inner_path->locator_type)) {
inner = inner_path;
} else {
/* Get inner bucketsize from cache which has saved before */
if (bucket->normal.nbuckets == virtualbuckets) {
thisbucketsize = bucket->normal.bucket_size;
*ndistinct = bucket->normal.ndistinct;
}
}
} else {
inner = inner_path;
if (((StreamPath*)inner)->type == STREAM_REDISTRIBUTE) {
if (bucket->redistribute.nbuckets == virtualbuckets) {
thisbucketsize = bucket->redistribute.bucket_size;
*ndistinct = bucket->redistribute.ndistinct;
}
} else if (((StreamPath*)inner)->type == STREAM_BROADCAST) {
if (bucket->broadcast.nbuckets == virtualbuckets) {
thisbucketsize = bucket->broadcast.bucket_size;
*ndistinct = bucket->broadcast.ndistinct;
}
}
}
/* Now we don't have cache for smp, should calculate every time */
if (inner_path->dop > 1) {
thisbucketsize = estimate_hash_bucketsize(root,
left ? get_leftop(restrictinfo->clause) : get_rightop(restrictinfo->clause),
virtualbuckets,
inner,
sjinfo,
ndistinct);
} else if (thisbucketsize < 0) {
/* not cached yet */
thisbucketsize = estimate_hash_bucketsize(root,
left ? get_leftop(restrictinfo->clause) : get_rightop(restrictinfo->clause),
virtualbuckets,
inner,
sjinfo,
ndistinct);
if (!IsA(inner_path, StreamPath)) {
if (!IsA(inner_path, HashPath)) {
bucket->normal.nbuckets = virtualbuckets;
bucket->normal.bucket_size = thisbucketsize;
bucket->normal.ndistinct = *ndistinct;
}
} else {
if (((StreamPath*)inner)->type == STREAM_REDISTRIBUTE) {
bucket->redistribute.nbuckets = virtualbuckets;
bucket->redistribute.bucket_size = thisbucketsize;
bucket->redistribute.ndistinct = *ndistinct;
} else if (((StreamPath*)inner)->type == STREAM_BROADCAST) {
bucket->broadcast.nbuckets = virtualbuckets;
bucket->broadcast.bucket_size = thisbucketsize;
bucket->broadcast.ndistinct = *ndistinct;
}
}
}
return thisbucketsize;
}
/*
* final_cost_hashjoin
* Final estimate of the cost and result size of a hashjoin path.
*
* Note: the numbatches estimate is also saved into 'path' for use later
*
* 'path' is already filled in except for the rows and cost fields and
* num_batches
* 'workspace' is the result from initial_cost_hashjoin
* 'sjinfo' is extra info about the join for selectivity estimation
* 'semifactors' contains valid data if path->jointype is SEMI or ANTI
*/
void final_cost_hashjoin(PlannerInfo* root, HashPath* path, JoinCostWorkspace* workspace, SpecialJoinInfo* sjinfo,
SemiAntiJoinFactors* semifactors, bool hasalternative, int dop)
{
Path* outer_path = path->jpath.outerjoinpath;
Path* inner_path = path->jpath.innerjoinpath;
double outer_path_rows = PATH_LOCAL_ROWS(outer_path) / dop;
double inner_path_rows = PATH_LOCAL_ROWS(inner_path) / dop;
double inner_path_rows_total = workspace->inner_rows_total;
List* hashclauses = path->path_hashclauses;
Cost startup_cost = workspace->startup_cost;
Cost run_cost = workspace->run_cost;
int numbuckets = workspace->numbuckets;
int numbatches = workspace->numbatches;
Cost cpu_per_tuple = 0.0;
QualCost hash_qual_cost;
QualCost qp_qual_cost;
double hashjointuples;
double virtualbuckets;
Selectivity innerbucketsize;
Selectivity outer_scan_ratio = 0.0;
ListCell* hcl = NULL;
ES_SELECTIVITY* es = NULL;
MemoryContext ExtendedStat;
MemoryContext oldcontext;
List* clauselist = hashclauses;
double innerdistinct = 1.0;
double tmp_distinct = 1.0;
double outerdistinct = 1.0;
/* Mark the path with the correct row estimate */
set_rel_path_rows(&path->jpath.path, path->jpath.path.parent, path->jpath.path.param_info);
/* For partial paths, scale row estimate. */
if (path->jpath.path.parallel_workers > 0) {
double parallel_divisor = get_parallel_divisor(&path->jpath.path);
path->jpath.path.rows = clamp_row_est(path->jpath.path.rows / parallel_divisor);
}
/*
* If inner_path or outer_path is EC functioinScan without stream,
* we should set the multiple particularly.
*/
set_joinpath_multiple_for_EC(root, &path->jpath.path, outer_path, inner_path);
/*
* We could include g_instance.cost_cxt.disable_cost in the preliminary estimate, but that
* would amount to optimizing for the case where the join method is
* disabled, which doesn't seem like the way to bet.
*/
if (!u_sess->attr.attr_sql.enable_hashjoin && hasalternative)
startup_cost += g_instance.cost_cxt.disable_cost;
/* mark the path with estimated # of batches */
path->num_batches = numbatches;
/* store the total number of tuples (sum of partial row estimates) */
path->inner_rows_total = inner_path_rows_total;
/* and compute the number of "virtual" buckets in the whole join */
virtualbuckets = (double)numbuckets * (double)numbatches;
/*
* Determine bucketsize fraction for inner relation. We use the smallest
* bucketsize estimated for any individual hashclause; this is undoubtedly
* conservative.
*
* BUT: if inner relation has been unique-ified, we can assume it's good
* for hashing. This is important both because it's the right answer, and
* because we avoid contaminating the cache with a value that's wrong for
* non-unique-ified paths.
*/
if (IsA(inner_path, UniquePath))
innerbucketsize = 1.0 / virtualbuckets;
else {
innerbucketsize = 1.0;
/* use extended statistics to calculate innerbucket size and outerbucketsize */
if (list_length(hashclauses) >= 2) {
ExtendedStat = AllocSetContextCreate(CurrentMemoryContext,
"ExtendedStat",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
oldcontext = MemoryContextSwitchTo(ExtendedStat);
es = New(ExtendedStat) ES_SELECTIVITY();
AssertEreport(root != NULL,
MOD_OPT,
"The NULL PlannerInfo is not allowed."
"when estimating the cost and result size of a hashjoin path.");
(void)es->calculate_selectivity(
root, hashclauses, sjinfo, path->jpath.jointype, &path->jpath, ES_COMPUTEBUCKETSIZE);
clauselist = es->unmatched_clause_group;
(void)MemoryContextSwitchTo(oldcontext);
ListCell* bcl = NULL;
Selectivity outerbucketsize = 1.0;
foreach (bcl, es->bucketsize_list) {
es_bucketsize* es_bucket = (es_bucketsize*)lfirst(bcl);
if (bms_is_subset(es_bucket->left_relids, inner_path->parent->relids)) {
innerbucketsize *=
es->estimate_hash_bucketsize(es_bucket, &tmp_distinct, true, inner_path, virtualbuckets);
innerdistinct *= tmp_distinct;
outerbucketsize *=
es->estimate_hash_bucketsize(es_bucket, &tmp_distinct, false, inner_path, virtualbuckets);
outerdistinct *= tmp_distinct;
} else {
AssertEreport(bms_is_subset(es_bucket->right_relids, inner_path->parent->relids),
MOD_OPT,
"The right relids is not subset of the relids of inner path's parent"
"when estimating the cost and result size of a hashjoin path.");
innerbucketsize *=
es->estimate_hash_bucketsize(es_bucket, &tmp_distinct, false, inner_path, virtualbuckets);
innerdistinct *= tmp_distinct;
outerbucketsize *=
es->estimate_hash_bucketsize(es_bucket, &tmp_distinct, true, inner_path, virtualbuckets);
outerdistinct *= tmp_distinct;
}
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("[ES]innerbucketsize: %e, innerdistinct: %.0f", innerbucketsize, innerdistinct)));
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("[ES]outerbucketsize: %e, outerdistinct: %.0f", outerbucketsize, outerdistinct)));
}
if (innerdistinct < MIN_HASH_BUCKET_SIZE) {
/* cut bucket size to min size, unless there are few distinct value in outer path */
outer_scan_ratio = Max(Min(innerdistinct / outerdistinct, 1.0), outer_scan_ratio);
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("[ES]outerdistinct: %f, outer_scan_ratio: %e", outerdistinct, outer_scan_ratio)));
} else
outer_scan_ratio = 1.0;
}
int number_of_joinrels = 0;
Bitmapset* join_relids = NULL;
foreach (hcl, clauselist) {
RestrictInfo* restrictinfo = (RestrictInfo*)lfirst(hcl);
Selectivity thisbucketsize;
Node* outerkey = NULL;
innerdistinct = 1.0;
AssertEreport(IsA(restrictinfo, RestrictInfo),
MOD_OPT,
"The nodeTag of restrictinfo is not T_RestrictInfo"
"when estimating the cost and result size of a hashjoin path.");
/*
* First we have to figure out which side of the hashjoin clause
* is the inner side.
*
* Since we tend to visit the same clauses over and over when
* planning a large query, we cache the bucketsize estimate in the
* RestrictInfo node to avoid repeated lookups of statistics.
*/
if (bms_is_subset(restrictinfo->right_relids, inner_path->parent->relids)) {
thisbucketsize =
compute_bucket_size(root, restrictinfo, virtualbuckets, inner_path, false, sjinfo, &innerdistinct);
outerkey = get_leftop(restrictinfo->clause);
} else {
AssertEreport(bms_is_subset(restrictinfo->left_relids, inner_path->parent->relids),
MOD_OPT,
"The left relids is not subset of the relids of inner path's parent"
"when estimating the cost and result size of a hashjoin path.");
thisbucketsize =
compute_bucket_size(root, restrictinfo, virtualbuckets, inner_path, true, sjinfo, &innerdistinct);
outerkey = get_rightop(restrictinfo->clause);
}
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("thisbucketsize: %e, innerdistinct: %.0f", thisbucketsize, innerdistinct)));
/*
* Adjust outer scan ratio if bucket size is too big, since we have at least 32768 buckets.
* If not all 32768 buckets are filled, the outer tuples can meet empty bucket and return
* immediately
*/
if (innerdistinct < MIN_HASH_BUCKET_SIZE) {
outerdistinct = DEFAULT_NUM_DISTINCT;
/*
* First estimate distinct value of outer path, will get derived stats for
* replicate path, hashjoin path or stream path
*/
if (IsA(outer_path, StreamPath) || IsA(outer_path, HashPath) ||
IsLocatorReplicated(outer_path->locator_type)) {
Selectivity outerbucketsize;
/* When calculating outerdistinct we have to take skew into consideration */
outerbucketsize =
estimate_hash_bucketsize(root, outerkey, virtualbuckets, outer_path, sjinfo, NULL);
/*
* Restrict outerdistinct less than MIN_HASH_BUCKET_SIZE
* otherwize outer_scan_ratio = innerdistinct / outerdistinct will be extremly small
* if outer rel have large amount of tuples and the cost will be estimated lower
* than it should be
*/
outerdistinct = Min(1 / outerbucketsize, MIN_HASH_BUCKET_SIZE);
}
/* cut bucket size to min size, unless there are few distinct value in outer path */
outer_scan_ratio = Max(Min(innerdistinct / outerdistinct, 1.0), outer_scan_ratio);
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("outerdistinct: %f, outer_scan_ratio: %e", outerdistinct, outer_scan_ratio)));
} else
outer_scan_ratio = 1.0;
ereport(DEBUG2, (errmodule(MOD_OPT_JOIN), errmsg("outer_scan_ratio: %e", outer_scan_ratio)));
/*
* Previously, we select the smallest innerbucketsize because there could be correlationship and the
* innerbucketsize could be too small if just multiply. For now, we have multi-column statistics and
* will calculate innerbucketsize first with multi-column statistics, code above.
*/
join_relids = bms_add_members(join_relids, restrictinfo->right_relids);
join_relids = bms_add_members(join_relids, restrictinfo->left_relids);
if (number_of_joinrels > 0 && number_of_joinrels == bms_num_members(join_relids)) {
/* There is no new rel added to the join_relids, which mean there could be some correlationship between
* clauses */
Selectivity tmp_bucketsize = innerbucketsize * thisbucketsize;
innerbucketsize = Min(innerbucketsize * 0.75, thisbucketsize * 0.75);
innerbucketsize = Max(innerbucketsize, tmp_bucketsize);
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("using fudge factor to fix innerbucket size: %e, tmp_bucketsize:%e",
innerbucketsize,
tmp_bucketsize)));
} else {
innerbucketsize *= thisbucketsize;
ereport(DEBUG2, (errmodule(MOD_OPT_JOIN), errmsg("multiplying innerbucket size: %e", innerbucketsize)));
}
number_of_joinrels = bms_num_members(join_relids);
}
if (join_relids != NULL) {
bms_free_ext(join_relids);
}
}
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("innerbucketsize: %e, outer_scan_ratio:%e", innerbucketsize, outer_scan_ratio)));
/*
* add some restrition for innerbucketsize:
* (1) innerbucketsize should not be smaller than 1.0e-7 as same as what we do in estimate_hash_bucketsize()
* (2) innerbucketsize * inner_path_rows should be more than 1.0 as there couldn't be that much distincts in single
* DN
*/
if (innerbucketsize * inner_path_rows < 1.0) {
innerbucketsize = 1 / clamp_row_est(inner_path_rows);
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("innerbucketsize: %e, inner_path_rows:%e", innerbucketsize, inner_path_rows)));
}
if (innerbucketsize < 1.0e-7) {
innerbucketsize = 1.0e-7;
ereport(DEBUG2, (errmodule(MOD_OPT_JOIN), errmsg("innerbucketsize: %e", innerbucketsize)));
}
path->jpath.path.innerdistinct = innerdistinct;
path->jpath.path.outerdistinct = outerdistinct;
/*
* Compute cost of the hashquals and qpquals (other restriction clauses)
* separately.
*/
cost_qual_eval(&hash_qual_cost, hashclauses, root);
cost_qual_eval(&qp_qual_cost, path->jpath.joinrestrictinfo, root);
qp_qual_cost.startup -= hash_qual_cost.startup;
qp_qual_cost.per_tuple -= hash_qual_cost.per_tuple;
/* CPU costs */
if (path->jpath.jointype == JOIN_SEMI || path->jpath.jointype == JOIN_ANTI) {
double outer_matched_rows;
Selectivity inner_scan_frac;
/*
* SEMI or ANTI join: executor will stop after first match.
*
* For an outer-rel row that has at least one match, we can expect the
* bucket scan to stop after a fraction 1/(match_count+1) of the
* bucket's rows, if the matches are evenly distributed. Since they
* probably aren't quite evenly distributed, we apply a fuzz factor of
* 2.0 to that fraction. (If we used a larger fuzz factor, we'd have
* to clamp inner_scan_frac to at most 1.0; but since match_count is
* at least 1, no such clamp is needed now.)
*/
outer_matched_rows = rint(outer_path_rows * semifactors->outer_match_frac);
inner_scan_frac = 2.0 / (semifactors->match_count + 1.0);
startup_cost += hash_qual_cost.startup;
double matching_cost = hash_qual_cost.per_tuple * clamp_row_est(outer_matched_rows * outer_scan_ratio) *
clamp_row_est(inner_path_rows * innerbucketsize * inner_scan_frac) * 0.5;
run_cost += matching_cost;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("matching_cost: %e, hash_qual_cost.per_tuple: %e, outer_matched_rows: %e, inner_scan_frac:%e",
matching_cost,
hash_qual_cost.per_tuple,
outer_matched_rows,
inner_scan_frac)));
/*
* For unmatched outer-rel rows, the picture is quite a lot different.
* In the first place, there is no reason to assume that these rows
* preferentially hit heavily-populated buckets; instead assume they
* are uncorrelated with the inner distribution and so they see an
* average bucket size of inner_path_rows / virtualbuckets. In the
* second place, it seems likely that they will have few if any exact
* hash-code matches and so very few of the tuples in the bucket will
* actually require eval of the hash quals. We don't have any good
* way to estimate how many will, but for the moment assume that the
* effective cost per bucket entry is one-tenth what it is for
* matchable tuples.
*/
run_cost += hash_qual_cost.per_tuple *
clamp_row_est((outer_path_rows - outer_matched_rows) * outer_scan_ratio) *
clamp_row_est(inner_path_rows / virtualbuckets) * 0.05;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add cpu cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
/* Get # of tuples that will pass the basic join */
if (path->jpath.jointype == JOIN_SEMI)
hashjointuples = outer_matched_rows;
else
hashjointuples = outer_path_rows - outer_matched_rows;
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("hashjointuples=%.0f, outer_tuples=%.0f, inner_tuples=%.0f, outer_global_tuples=%.0f, "
"inner_global_tuples=%.0f, outer_matched_rows=%.10f",
hashjointuples,
outer_path_rows,
inner_path_rows,
outer_path->rows,
inner_path->rows,
outer_matched_rows)));
} else if (path->jpath.jointype == JOIN_RIGHT_SEMI || path->jpath.jointype == JOIN_RIGHT_ANTI) {
double outer_matched_rows, inner_matched_rows;
/*
* RIGHT_SEMI or RIGHT_ANTI join: hash table is the side to be returned
*
* For cost of this Join, we should be careful that the second member (i.e. match_count)
* of structure SemiAntiJoinFactors means the fraction of the inner tuples that are
* expected to have at least one match in outer side. For this new meaning, we
* have two selectivites:
* outer_match_frac: percent of outer-rel rows which have at least one match
* inner_match_frac: percent of inner-rel rows which have at least one match
*
* Note:
* 1. the caller should carry semifactors with the new meaning for this 'RIGHT' join.
* 2. outer_match_rows should be constrained to a limit by distinct value:
* if: outer rows: N, with distinct d2; inner rows: with distinct d1
* outer rows which have matches: N'
* then,
* N'
* --- * d2 <= d1, (averagely, without data skew)
* N
* Hence, N' should be corrected, e.g. N' = min (N', N * d1/d2)
*/
outer_matched_rows = rint(outer_path_rows * semifactors->outer_match_frac);
outer_matched_rows = Min(outer_matched_rows, outer_scan_ratio * outer_path_rows);
inner_matched_rows = rint(inner_path_rows * semifactors->match_count);
startup_cost += hash_qual_cost.startup;
/*
* For matched outer-rel rows, the idea is:
* Let
* N=outer_path_rows, N'=outer_matched_rows
* n=inner_path_rows, n'=inner_matched_rows.
* Since N' rows of outer-rel match n' rows of inner-rel, 1 row of outer-rel roughly
* matches n'/N' rows of inner-rel. Suppose that the row in Hash Table is deleted
* once it is matched with outer-rel. In this case, 1st match row of outer-rel will
* see a Hash Table holding n rows, 2nd match row of outer-rel will see n - n'/N'
* rows in the Hash Table, the 3td will see n - 2 * n'/N' rows, and so on.
* Recall that, during probe stage, one outer row do not stop even if it matches
* an inner row in Hash Table, therefore, these outer-rel rows (can match sth in
* Hash Table) shall do averagely
* innerbucketsize * ( n + (n - n'/N') + (n - 2n'/N') + ... + (n - (N'-1)n'/N') )
* = innerbucketsize * ( n*N' - n'*(N'-1)/2 )
* comparisons totaly.
*/
run_cost += hash_qual_cost.per_tuple *
clamp_row_est(innerbucketsize * (inner_path_rows * outer_matched_rows -
0.5 * inner_matched_rows * (outer_matched_rows - 1.0)));
/* cost of moving pointer which used to delete cells in Hash table */
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * 0.1 * outer_matched_rows *
clamp_row_est((inner_path_rows - inner_matched_rows * 0.5) / virtualbuckets);
/*
* For unmatched outer-rel rows:
* By statistical average, each unmatched row will see (n - n'/2) rows and compare
* an bucket with size being of '(n - n'/2)/virtualbuckets' if it could have exact hash-
* code matches. But, it is most likely that very few(say 1/10) of these unmatched
* outer rows can exactly match inner rows. That is only these '1/10' will actually
* require evaluation, while others '9/10' match empty buckets.
*/
run_cost += hash_qual_cost.per_tuple * (outer_path_rows - outer_matched_rows) *
clamp_row_est((inner_path_rows - inner_matched_rows * 0.5) / virtualbuckets) * 0.05;
/* cost of moving pointer which used to delete cells in Hash table: assume 1/10 needs compare */
run_cost += u_sess->attr.attr_sql.cpu_operator_cost * 0.1 * (outer_path_rows - outer_matched_rows) * 0.05 *
clamp_row_est((inner_path_rows - inner_matched_rows * 0.5) / virtualbuckets);
if (path->jpath.jointype == JOIN_RIGHT_SEMI)
hashjointuples = inner_matched_rows;
else {
/* hashjointuples are those unmatched tuples in hash table */
hashjointuples = inner_path_rows - inner_matched_rows;
}
/* log messages for debug */
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add cpu cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("hashjointuples=%.0f, outer_tuples=%.0f, inner_tuples=%.0f, outer_global_tuples=%.0f, "
"inner_global_tuples=%.0f, outer_matched_rows=%.10f, inner_matched_rows=%.10f",
hashjointuples,
outer_path_rows,
inner_path_rows,
outer_path->rows,
inner_path->rows,
outer_matched_rows,
inner_matched_rows)));
} else {
/*
* The number of tuple comparisons needed is the number of outer
* tuples times the typical number of tuples in a hash bucket, which
* is the inner relation size times its bucketsize fraction. At each
* one, we need to evaluate the hashjoin quals. But actually,
* charging the full qual eval cost at each tuple is pessimistic,
* since we don't evaluate the quals unless the hash values match
* exactly. For lack of a better idea, halve the cost estimate to
* allow for that.
*/
startup_cost += hash_qual_cost.startup;
run_cost += hash_qual_cost.per_tuple * clamp_row_est(outer_path_rows * outer_scan_ratio) *
clamp_row_est(inner_path_rows * innerbucketsize) * 0.5;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add cpu cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
/*
* Get approx # tuples passing the hashquals. We use
* approx_tuple_count here because we need an estimate done with
* JOIN_INNER semantics.
*/
hashjointuples = approx_tuple_count(root, &path->jpath, hashclauses);
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("outer_path_rows=%.0f, inner_path_rows=%.0f, hashjointuples=%.0f",
outer_path_rows,
inner_path_rows,
hashjointuples)));
}
/*
* For each tuple that gets through the hashjoin proper, we charge
* u_sess->attr.attr_sql.cpu_tuple_cost plus the cost of evaluating additional restriction
* clauses that are to be applied at the join. (This is pessimistic since
* not all of the quals may get evaluated at each tuple.)
*/
startup_cost += qp_qual_cost.startup;
cpu_per_tuple = u_sess->attr.attr_sql.cpu_tuple_cost + qp_qual_cost.per_tuple;
run_cost += cpu_per_tuple * hashjointuples;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add restriction clauses cost: startup_cost: %lf, run_cost: %lf", startup_cost, run_cost)));
path->jpath.path.startup_cost = startup_cost;
path->jpath.path.total_cost = startup_cost + run_cost;
path->jpath.path.stream_cost = inner_path->stream_cost;
if (!u_sess->attr.attr_sql.enable_hashjoin && hasalternative)
path->jpath.path.total_cost *= g_instance.cost_cxt.disable_cost_enlarge_factor;
ereport(DEBUG2,
(errmodule(MOD_OPT_JOIN),
errmsg("Add cpu cost: stream_cost: %lf, startup_cost: %lf, total_cost: %lf",
path->jpath.path.stream_cost,
path->jpath.path.startup_cost,
path->jpath.path.total_cost)));
copy_mem_info(&path->mem_info, &workspace->inner_mem_info);
debug_print_hashjoin_detail(
root, path, virtualbuckets, innerbucketsize, outer_scan_ratio, startup_cost, startup_cost + run_cost);
/* free space used by extended statistic */
if (es != NULL) {
clauselist = NIL;
list_free_ext(es->unmatched_clause_group);
delete es;
MemoryContextDelete(ExtendedStat);
}
}
/*
* cost_subplan
* Figure the costs for a SubPlan (or initplan).
*
* Note: we could dig the subplan's Plan out of the root list, but in practice
* all callers have it handy already, so we make them pass it.
*/
void cost_subplan(PlannerInfo* root, SubPlan* subplan, Plan* plan)
{
QualCost sp_cost;
/* Figure any cost for evaluating the testexpr */
cost_qual_eval(&sp_cost, make_ands_implicit((Expr*)subplan->testexpr), root);
if (subplan->useHashTable) {
/*
* If we are using a hash table for the subquery outputs, then the
* cost of evaluating the query is a one-time cost. We charge one
* u_sess->attr.attr_sql.cpu_operator_cost per tuple for the work of loading the hashtable,
* too.
*/
sp_cost.startup += plan->total_cost + u_sess->attr.attr_sql.cpu_operator_cost * PLAN_LOCAL_ROWS(plan);
/*
* The per-tuple costs include the cost of evaluating the lefthand
* expressions, plus the cost of probing the hashtable. We already
* accounted for the lefthand expressions as part of the testexpr, and
* will also have counted one u_sess->attr.attr_sql.cpu_operator_cost for each comparison
* operator. That is probably too low for the probing cost, but it's
* hard to make a better estimate, so live with it for now.
*/
} else {
/*
* Otherwise we will be rescanning the subplan output on each
* evaluation. We need to estimate how much of the output we will
* actually need to scan. NOTE: this logic should agree with the
* tuple_fraction estimates used by make_subplan() in
* plan/subselect.c.
*/
Cost plan_run_cost = plan->total_cost - plan->startup_cost;
if (subplan->subLinkType == EXISTS_SUBLINK) {
/* we only need to fetch 1 tuple */
sp_cost.per_tuple += plan_run_cost / PLAN_LOCAL_ROWS(plan);
} else if (subplan->subLinkType == ALL_SUBLINK || subplan->subLinkType == ANY_SUBLINK) {
/* assume we need 50% of the tuples */
sp_cost.per_tuple += 0.50 * plan_run_cost;
/* also charge a u_sess->attr.attr_sql.cpu_operator_cost per row examined */
sp_cost.per_tuple += 0.50 * PLAN_LOCAL_ROWS(plan) * u_sess->attr.attr_sql.cpu_operator_cost;
} else {
/* assume we need all tuples */
sp_cost.per_tuple += plan_run_cost;
}
/*
* Also account for subplan's startup cost. If the subplan is
* uncorrelated or undirect correlated, AND its topmost node is one
* that materializes its output, assume that we'll only need to pay
* its startup cost once; otherwise assume we pay the startup cost
* every time.
*/
if (subplan->parParam == NIL && ExecMaterializesOutput(nodeTag(plan)))
sp_cost.startup += plan->startup_cost;
else
sp_cost.per_tuple += plan->startup_cost;
}
subplan->startup_cost = sp_cost.startup;
subplan->per_call_cost = sp_cost.per_tuple;
}
/*
* cost_rescan
* Given a finished Path, estimate the costs of rescanning it after
* having done so the first time. For some Path types a rescan is
* cheaper than an original scan (if no parameters change), and this
* function embodies knowledge about that. The default is to return
* the same costs stored in the Path. (Note that the cost estimates
* actually stored in Paths are always for first scans.)
*
* This function is not currently intended to model effects such as rescans
* being cheaper due to disk block caching; what we are concerned with is
* plan types wherein the executor caches results explicitly, or doesn't
* redo startup calculations, etc.
*/
void cost_rescan(PlannerInfo* root, Path* path, Cost* rescan_startup_cost, /* output parameters */
Cost* rescan_total_cost, OpMemInfo* mem_info)
{
int dop = SET_DOP(path->dop);
switch (path->pathtype) {
case T_FunctionScan:
/*
* Currently, nodeFunctionscan.c always executes the function to
* completion before returning any rows, and caches the results in
* a tuplestore. So the function eval cost is all startup cost
* and isn't paid over again on rescans. However, all run costs
* will be paid over again.
*/
*rescan_startup_cost = 0;
*rescan_total_cost = path->total_cost - path->startup_cost;
break;
case T_HashJoin:
/*
* Assume that all of the startup cost represents hash table
* building, which we won't have to do over.
*/
*rescan_startup_cost = 0;
*rescan_total_cost = path->total_cost - path->startup_cost;
break;
case T_CteScan:
case T_WorkTableScan: {
/*
* These plan types materialize their final result in a
* tuplestore or tuplesort object. So the rescan cost is only
* u_sess->attr.attr_sql.cpu_tuple_cost per tuple, unless the result is large enough
* to spill to disk.
*/
double rows = PATH_LOCAL_ROWS(path);
Cost run_cost = u_sess->attr.attr_sql.cpu_tuple_cost * rows;
double nbytes = relation_byte_size(rows, path->parent->width, false, true, false);
long work_mem_bytes = u_sess->opt_cxt.op_work_mem * 1024L;
if (nbytes > work_mem_bytes) {
/* It will spill, so account for re-read cost */
double npages = ceil(nbytes / BLCKSZ);
run_cost += u_sess->attr.attr_sql.seq_page_cost * npages;
}
*rescan_startup_cost = 0;
*rescan_total_cost = run_cost;
} break;
case T_Material:
case T_Sort: {
/*
* These plan types not only materialize their results, but do
* not implement qual filtering or projection. So they are
* even cheaper to rescan than the ones above. We charge only
* cpu_operator_cost per tuple. (Note: keep that in sync with
* the run_cost charge in cost_sort, and also see comments in
* cost_material before you change it.)
*/
double rows = PATH_LOCAL_ROWS(path);
*rescan_startup_cost = 0;
*rescan_total_cost = cost_rescan_material(rows,
get_path_actual_total_width(path, root->glob->vectorized, OP_MATERIAL),
mem_info,
root->glob->vectorized,
dop);
} break;
default:
*rescan_startup_cost = path->startup_cost;
*rescan_total_cost = path->total_cost;
break;
}
}
/*
* cost_rescan_material
* Calculate rescan cost of the materialize node
*
* Parameters:
* @in rows: number of rows of input relation
* @in width: width of input relation
* @out mem_info: mem info to record memory usage of materialize node
* @in vectorized: if path to be vectorized plan
*
* Returns: calculated cost
*/
Cost cost_rescan_material(double rows, int width, OpMemInfo* mem_info, bool vectorized, int dop)
{
/*
* These plan types not only materialize their results, but do
* not implement qual filtering or projection. So they are
* even cheaper to rescan than the ones above. We charge only
* cpu_operator_cost per tuple. (Note: keep that in sync with
* the run_cost charge in cost_sort, and also see comments in
* cost_material before you change it.)
*/
double local_rows = rows / dop;
Cost run_cost = u_sess->attr.attr_sql.cpu_operator_cost * local_rows;
double nbytes = relation_byte_size(local_rows, width, vectorized, true, false);
long work_mem_bytes = u_sess->opt_cxt.op_work_mem * 1024L / dop;
/* It will spill, so account for re-read cost */
double npages = ceil(nbytes / BLCKSZ);
double disk_cost = u_sess->attr.attr_sql.seq_page_cost * npages;
if (nbytes > work_mem_bytes) {
run_cost += disk_cost;
}
if (mem_info != NULL) {
mem_info->opMem = u_sess->opt_cxt.op_work_mem;
mem_info->maxMem = nbytes / 1024L * dop;
mem_info->minMem = mem_info->maxMem / SORT_MAX_DISK_SIZE;
mem_info->regressCost = disk_cost;
}
return run_cost;
}
#ifdef PGXC
/*
* cost_remotequery
* As of now the function just sets the costs to 0 to make this path the
* cheapest.
* NOTICE: Ideally, we should estimate the costs of network transfer from
* datanodes and any datanode costs involved.
*/
void cost_remotequery(RemoteQueryPath* rqpath, PlannerInfo* root, RelOptInfo* rel)
{
rqpath->path.startup_cost = 0;
rqpath->path.total_cost = 0;
set_rel_path_rows(&rqpath->path, rel, NULL);
}
#endif /* PGXC */
/*
* cost_qual_eval
* Estimate the CPU costs of evaluating a WHERE clause.
* The input can be either an implicitly-ANDed list of boolean
* expressions, or a list of RestrictInfo nodes. (The latter is
* preferred since it allows caching of the results.)
* The result includes both a one-time (startup) component,
* and a per-evaluation component.
*/
void cost_qual_eval(QualCost* cost, List* quals, PlannerInfo* root)
{
cost_qual_eval_context context;
ListCell* l = NULL;
context.root = root;
context.total.startup = 0;
context.total.per_tuple = 0;
/* We don't charge any cost for the implicit ANDing at top level ... */
foreach (l, quals) {
Node* qual = (Node*)lfirst(l);
(void)cost_qual_eval_walker(qual, &context);
}
*cost = context.total;
}
/*
* cost_qual_eval_node
* As above, for a single RestrictInfo or expression.
*/
void cost_qual_eval_node(QualCost* cost, Node* qual, PlannerInfo* root)
{
cost_qual_eval_context context;
context.root = root;
context.total.startup = 0;
context.total.per_tuple = 0;
(void)cost_qual_eval_walker(qual, &context);
*cost = context.total;
}
static bool cost_qual_eval_walker(Node* node, cost_qual_eval_context* context)
{
if (node == NULL)
return false;
/*
* RestrictInfo nodes contain an eval_cost field reserved for this
* routine's use, so that it's not necessary to evaluate the qual clause's
* cost more than once. If the clause's cost hasn't been computed yet,
* the field's startup value will contain -1.
*/
if (IsA(node, RestrictInfo)) {
RestrictInfo* rinfo = (RestrictInfo*)node;
if (rinfo->eval_cost.startup < 0) {
cost_qual_eval_context locContext;
locContext.root = context->root;
locContext.total.startup = 0;
locContext.total.per_tuple = 0;
/*
* For an OR clause, recurse into the marked-up tree so that we
* set the eval_cost for contained RestrictInfos too.
*/
if (rinfo->orclause)
(void)cost_qual_eval_walker((Node*)rinfo->orclause, &locContext);
else
(void)cost_qual_eval_walker((Node*)rinfo->clause, &locContext);
/*
* If the RestrictInfo is marked pseudoconstant, it will be tested
* only once, so treat its cost as all startup cost.
*/
if (rinfo->pseudoconstant) {
/* count one execution during startup */
locContext.total.startup += locContext.total.per_tuple;
locContext.total.per_tuple = 0;
}
rinfo->eval_cost = locContext.total;
}
context->total.startup += rinfo->eval_cost.startup;
context->total.per_tuple += rinfo->eval_cost.per_tuple;
/* do NOT recurse into children */
return false;
}
/*
* For each operator or function node in the given tree, we charge the
* estimated execution cost given by pg_proc.procost (remember to multiply
* this by cpu_operator_cost).
*
* Vars and Consts are charged zero, and so are boolean operators (AND,
* OR, NOT). Simplistic, but a lot better than no model at all.
*
* Should we try to account for the possibility of short-circuit
* evaluation of AND/OR? Probably *not*, because that would make the
* results depend on the clause ordering, and we are not in any position
* to expect that the current ordering of the clauses is the one that's
* going to end up being used. The above per-RestrictInfo caching would
* not mix well with trying to re-order clauses anyway.
*
* Another issue that is entirely ignored here is that if a set-returning
* function is below top level in the tree, the functions/operators above
* it will need to be evaluated multiple times. In practical use, such
* cases arise so seldom as to not be worth the added complexity needed;
* moreover, since our rowcount estimates for functions tend to be pretty
* phony, the results would also be pretty phony.
*/
if (IsA(node, FuncExpr)) {
context->total.per_tuple += get_func_cost(((FuncExpr*)node)->funcid) * u_sess->attr.attr_sql.cpu_operator_cost;
} else if (IsA(node, OpExpr) || IsA(node, DistinctExpr) || IsA(node, NullIfExpr)) {
/* rely on struct equivalence to treat these all alike */
set_opfuncid((OpExpr*)node);
context->total.per_tuple += get_func_cost(((OpExpr*)node)->opfuncid) * u_sess->attr.attr_sql.cpu_operator_cost;
} else if (IsA(node, ScalarArrayOpExpr)) {
/*
* Estimate that the operator will be applied to about half of the
* array elements before the answer is determined.
*/
ScalarArrayOpExpr* saop = (ScalarArrayOpExpr*)node;
Node* arraynode = (Node*)lsecond(saop->args);
set_sa_opfuncid(saop);
context->total.per_tuple += get_func_cost(saop->opfuncid) * u_sess->attr.attr_sql.cpu_operator_cost *
estimate_array_length(arraynode) * 0.5;
} else if (IsA(node, Aggref) || IsA(node, WindowFunc)) {
/*
* Aggref and WindowFunc nodes are (and should be) treated like Vars,
* ie, zero execution cost in the current model, because they behave
* essentially like Vars in execQual.c. We disregard the costs of
* their input expressions for the same reason. The actual execution
* costs of the aggregate/window functions and their arguments have to
* be factored into plan-node-specific costing of the Agg or WindowAgg
* plan node.
*/
return false; /* don't recurse into children */
} else if (IsA(node, CoerceViaIO)) {
CoerceViaIO* iocoerce = (CoerceViaIO*)node;
Oid iofunc;
Oid typioparam;
bool typisvarlena = false;
/* check the result type's input function */
getTypeInputInfo(iocoerce->resulttype, &iofunc, &typioparam);
context->total.per_tuple += get_func_cost(iofunc) * u_sess->attr.attr_sql.cpu_operator_cost;
/* check the input type's output function */
getTypeOutputInfo(exprType((Node*)iocoerce->arg), &iofunc, &typisvarlena);
context->total.per_tuple += get_func_cost(iofunc) * u_sess->attr.attr_sql.cpu_operator_cost;
} else if (IsA(node, ArrayCoerceExpr)) {
ArrayCoerceExpr* acoerce = (ArrayCoerceExpr*)node;
Node* arraynode = (Node*)acoerce->arg;
if (OidIsValid(acoerce->elemfuncid))
context->total.per_tuple += get_func_cost(acoerce->elemfuncid) * u_sess->attr.attr_sql.cpu_operator_cost *
estimate_array_length(arraynode);
} else if (IsA(node, RowCompareExpr)) {
/* Conservatively assume we will check all the columns */
RowCompareExpr* rcexpr = (RowCompareExpr*)node;
ListCell* lc = NULL;
foreach (lc, rcexpr->opnos) {
Oid opid = lfirst_oid(lc);
context->total.per_tuple += get_func_cost(get_opcode(opid)) * u_sess->attr.attr_sql.cpu_operator_cost;
}
} else if (IsA(node, CurrentOfExpr)) {
/* Report high cost to prevent selection of anything but TID scan */
context->total.startup += g_instance.cost_cxt.disable_cost;
} else if (IsA(node, SubLink)) {
/* This routine should not be applied to un-planned expressions */
ereport(ERROR,
(errmodule(MOD_OPT),
errcode(ERRCODE_OPTIMIZER_INCONSISTENT_STATE),
errmsg("cannot handle unplanned sub-select when costing quals")));
} else if (IsA(node, SubPlan)) {
/*
* A subplan node in an expression typically indicates that the
* subplan will be executed on each evaluation, so charge accordingly.
* (Sub-selects that can be executed as InitPlans have already been
* removed from the expression.)
*/
SubPlan* subplan = (SubPlan*)node;
context->total.startup += subplan->startup_cost;
context->total.per_tuple += subplan->per_call_cost;
/*
* We don't want to recurse into the testexpr, because it was already
* counted in the SubPlan node's costs. So we're done.
*/
return false;
} else if (IsA(node, AlternativeSubPlan)) {
/*
* Arbitrarily use the first alternative plan for costing. (We should
* certainly only include one alternative, and we don't yet have
* enough information to know which one the executor is most likely to
* use.)
*/
AlternativeSubPlan* asplan = (AlternativeSubPlan*)node;
return cost_qual_eval_walker((Node*)linitial(asplan->subplans), context);
}
/* recurse into children */
return expression_tree_walker(node, (bool (*)())cost_qual_eval_walker, (void*)context);
}
/*
* get_restriction_qual_cost
* Compute evaluation costs of a baserel's restriction quals, plus any
* movable join quals that have been pushed down to the scan.
* Results are returned into *qpqual_cost.
*
* This is a convenience subroutine that works for seqscans and other cases
* where all the given quals will be evaluated the hard way. It's not useful
* for cost_index(), for example, where the index machinery takes care of
* some of the quals. We assume baserestrictcost was previously set
* by set_baserel_size_estimates().
*/
static void get_restriction_qual_cost(
PlannerInfo* root, RelOptInfo* baserel, ParamPathInfo* param_info, QualCost* qpqual_cost)
{
if (param_info != NULL) {
/* Include costs of pushed-down clauses */
cost_qual_eval(qpqual_cost, param_info->ppi_clauses, root);
qpqual_cost->startup += baserel->baserestrictcost.startup;
qpqual_cost->per_tuple += baserel->baserestrictcost.per_tuple;
} else
*qpqual_cost = baserel->baserestrictcost;
}
/*
* compute_semi_anti_join_factors
* Estimate how much of the inner input a SEMI or ANTI join
* can be expected to scan.
*
* In a hash or nestloop SEMI/ANTI join, the executor will stop scanning
* inner rows as soon as it finds a match to the current outer row.
* We should therefore adjust some of the cost components for this effect.
* This function computes some estimates needed for these adjustments.
* These estimates will be the same regardless of the particular paths used
* for the outer and inner relation, so we compute these once and then pass
* them to all the join cost estimation functions.
*
* Input parameters:
* outerrel: outer relation under consideration
* innerrel: inner relation under consideration
* jointype: must be JOIN_SEMI or JOIN_ANTI
* sjinfo: SpecialJoinInfo relevant to this join
* restrictlist: join quals
* Output parameters:
* *semifactors is filled in (see relation.h for field definitions)
*/
void compute_semi_anti_join_factors(PlannerInfo* root, RelOptInfo* outerrel, RelOptInfo* innerrel, JoinType jointype,
SpecialJoinInfo* sjinfo, List* restrictlist, SemiAntiJoinFactors* semifactors)
{
Selectivity jselec;
Selectivity nselec;
Selectivity avgmatch;
SpecialJoinInfo norm_sjinfo;
List* joinquals = NIL;
ListCell* l = NULL;
/* Should only be called in these cases */
AssertEreport(jointype == JOIN_SEMI || jointype == JOIN_ANTI,
MOD_OPT,
"Only JOIN_SEMI or JOIN_ANTI can be supported"
"when estimating how much of the inner input a SEMI or ANTI join can be expected to scan.");
/*
* In an ANTI join, we must ignore clauses that are "pushed down", since
* those won't affect the match logic. In a SEMI join, we do not
* distinguish joinquals from "pushed down" quals, so just use the whole
* restrictinfo list.
*/
if (jointype == JOIN_ANTI) {
joinquals = NIL;
foreach (l, restrictlist) {
RestrictInfo* rinfo = (RestrictInfo*)lfirst(l);
AssertEreport(IsA(rinfo, RestrictInfo),
MOD_OPT,
"The nodeTag of rinfo is T_RestrictInfo"
"when estimating how much of the inner input a SEMI or ANTI join can be expected to scan.");
if (!rinfo->is_pushed_down)
joinquals = lappend(joinquals, rinfo);
}
} else
joinquals = restrictlist;
/*
* Get the JOIN_SEMI or JOIN_ANTI selectivity of the join clauses.
*/
jselec = clauselist_selectivity(root, joinquals, 0, jointype, sjinfo);
/*
* Also get the normal inner-join selectivity of the join clauses.
*/
norm_sjinfo.type = T_SpecialJoinInfo;
norm_sjinfo.min_lefthand = outerrel->relids;
norm_sjinfo.min_righthand = innerrel->relids;
norm_sjinfo.syn_lefthand = outerrel->relids;
norm_sjinfo.syn_righthand = innerrel->relids;
norm_sjinfo.jointype = JOIN_INNER;
/* we don't bother trying to make the remaining fields valid */
norm_sjinfo.lhs_strict = false;
norm_sjinfo.delay_upper_joins = false;
norm_sjinfo.join_quals = NIL;
/* we don't cache nselec because it is only used to compute matched cout for inner. */
norm_sjinfo.varratio_cached = false;
nselec = clauselist_selectivity(root, joinquals, 0, JOIN_INNER, &norm_sjinfo);
/* Avoid leaking a lot of ListCells */
if (jointype == JOIN_ANTI)
list_free_ext(joinquals);
/*
* jselec can be interpreted as the fraction of outer-rel rows that have
* any matches (this is true for both SEMI and ANTI cases). And nselec is
* the fraction of the Cartesian product that matches. So, the average
* number of matches for each outer-rel row that has at least one match is
* nselec * inner_rows / jselec.
*
* Note: it is correct to use the inner rel's "rows" count here, even
* though we might later be considering a parameterized inner path with
* fewer rows. This is because we have included all the join clauses in
* the selectivity estimate.
*/
if (jselec > 0) { /* protect against zero divide */
avgmatch = nselec * RELOPTINFO_LOCAL_FIELD(root, innerrel, rows) / jselec;
/* Clamp to sane range */
avgmatch = Max(1.0, avgmatch);
} else {
avgmatch = 1.0;
}
semifactors->outer_match_frac = jselec;
semifactors->match_count = avgmatch;
}
/*
* has_indexed_join_quals
* Check whether all the joinquals of a nestloop join are used as
* inner index quals.
*
* If the inner path of a SEMI/ANTI join is an indexscan (including bitmap
* indexscan) that uses all the joinquals as indexquals, we can assume that an
* unmatched outer tuple is cheap to process, whereas otherwise it's probably
* expensive.
*/
bool has_indexed_join_quals(NestPath* joinpath)
{
Relids joinrelids = joinpath->path.parent->relids;
Path* innerpath = joinpath->innerjoinpath;
List* indexclauses = NIL;
bool found_one = false;
ListCell* lc = NULL;
/* If join still has quals to evaluate, it's not fast */
if (joinpath->joinrestrictinfo != NIL)
return false;
/* Nor if the inner path isn't parameterized at all */
if (innerpath->param_info == NULL)
return false;
/* Find the indexclauses list for the inner scan */
switch (innerpath->pathtype) {
case T_IndexScan:
case T_IndexOnlyScan:
indexclauses = ((IndexPath*)innerpath)->indexclauses;
break;
case T_BitmapHeapScan: {
/* Accept only a simple bitmap scan, not AND/OR cases */
Path* bmqual = ((BitmapHeapPath*)innerpath)->bitmapqual;
if (IsA(bmqual, IndexPath))
indexclauses = ((IndexPath*)bmqual)->indexclauses;
else
return false;
break;
}
default:
/*
* If it's not a simple indexscan, it probably doesn't run quickly
* for zero rows out, even if it's a parameterized path using all
* the joinquals.
*/
return false;
}
/*
* Examine the inner path's param clauses. Any that are from the outer
* path must be found in the indexclauses list, either exactly or in an
* equivalent form generated by equivclass.c. Also, we must find at least
* one such clause, else it's a clauseless join which isn't fast.
*/
found_one = false;
foreach (lc, innerpath->param_info->ppi_clauses) {
RestrictInfo* rinfo = (RestrictInfo*)lfirst(lc);
if (join_clause_is_movable_into(rinfo, innerpath->parent->relids, joinrelids)) {
if (!(list_member_ptr(indexclauses, rinfo) || is_redundant_derived_clause(rinfo, indexclauses)))
return false;
found_one = true;
}
}
return found_one;
}
/*
* approx_tuple_count
* Quick-and-dirty estimation of the number of join rows passing
* a set of qual conditions.
*
* The quals can be either an implicitly-ANDed list of boolean expressions,
* or a list of RestrictInfo nodes (typically the latter).
*
* We intentionally compute the selectivity under JOIN_INNER rules, even
* if it's some type of outer join. This is appropriate because we are
* trying to figure out how many tuples pass the initial merge or hash
* join step.
*
* This is quick-and-dirty because we bypass clauselist_selectivity, and
* simply multiply the independent clause selectivities together. Now
* clauselist_selectivity often can't do any better than that anyhow, but
* for some situations (such as range constraints) it is smarter. However,
* we can't effectively cache the results of clauselist_selectivity, whereas
* the individual clause selectivities can be and are cached.
*
* Since we are only using the results to estimate how many potential
* output tuples are generated and passed through qpqual checking, it
* seems OK to live with the approximation.
*/
double approx_tuple_count(PlannerInfo* root, JoinPath* path, List* quals)
{
double tuples;
double outer_global_tuples = path->outerjoinpath->rows;
double inner_global_tuples = path->innerjoinpath->rows;
double outer_tuples = PATH_LOCAL_ROWS(path->outerjoinpath);
double inner_tuples = PATH_LOCAL_ROWS(path->innerjoinpath);
SpecialJoinInfo sjinfo;
Selectivity selec = 1.0;
ListCell* l = NULL;
int dop = SET_DOP(path->path.dop);
List* qual_list = quals;
ES_SELECTIVITY* es = NULL;
MemoryContext ExtendedStat;
MemoryContext oldcontext;
/*
* Make up a SpecialJoinInfo for JOIN_INNER semantics.
*/
sjinfo.type = T_SpecialJoinInfo;
sjinfo.min_lefthand = path->outerjoinpath->parent->relids;
sjinfo.min_righthand = path->innerjoinpath->parent->relids;
sjinfo.syn_lefthand = path->outerjoinpath->parent->relids;
sjinfo.syn_righthand = path->innerjoinpath->parent->relids;
sjinfo.jointype = JOIN_INNER;
/* we don't bother trying to make the remaining fields valid */
sjinfo.lhs_strict = false;
sjinfo.delay_upper_joins = false;
sjinfo.join_quals = NIL;
sjinfo.varratio_cached = true;
/* initialize es_selectivity class */
if (list_length(qual_list) >= 2) {
ExtendedStat = AllocSetContextCreate(CurrentMemoryContext,
"ExtendedStat",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
oldcontext = MemoryContextSwitchTo(ExtendedStat);
es = New(ExtendedStat) ES_SELECTIVITY();
AssertEreport(root != NULL,
MOD_OPT,
"The NULL PlannerInfo is not allowed "
"when estimation the number of join rows passing a set of qual conditions approximately.");
selec = es->calculate_selectivity(root, qual_list, &sjinfo, JOIN_INNER, path, ES_EQJOINSEL);
qual_list = es->unmatched_clause_group;
(void)MemoryContextSwitchTo(oldcontext);
}
/* Get the approximate selectivity */
foreach (l, qual_list) {
Node* qual = (Node*)lfirst(l);
/* Note that clause_selectivity will be able to cache its result */
selec *= clause_selectivity(root, qual, 0, JOIN_INNER, &sjinfo);
}
/* We should use global tuples if has stream. */
if (IS_STREAM_PLAN) {
/*
* if outerpath or innerpath is redistribute, we should use global tuples for both.
* if outerpath or innerpath is broadcast, we should use local tuples for the side of broadcast.
* if outerpath or innerpath is not stream, we should use global tuples for both.
* When parallel broadcast or local broadcast, the inner_tuples refer to tuples in a DN, so we divide dop.
*/
if (IsA(path->outerjoinpath, StreamPath) && (STREAM_BROADCAST == ((StreamPath*)path->outerjoinpath)->type)) {
outer_global_tuples = outer_tuples / dop;
} else if (IsA(path->innerjoinpath, StreamPath) &&
(((StreamPath*)path->innerjoinpath)->type == STREAM_BROADCAST)) {
inner_global_tuples = inner_tuples / dop;
} else if (IsA(path->outerjoinpath, StreamPath) && ((StreamPath*)path->outerjoinpath)->smpDesc &&
((StreamPath*)path->outerjoinpath)->smpDesc->distriType == LOCAL_BROADCAST) {
outer_global_tuples = outer_global_tuples / dop;
} else if (IsA(path->innerjoinpath, StreamPath) && ((StreamPath*)path->innerjoinpath)->smpDesc &&
((StreamPath*)path->innerjoinpath)->smpDesc->distriType == LOCAL_BROADCAST) {
inner_global_tuples = inner_global_tuples / dop;
}
tuples = selec * outer_global_tuples * inner_global_tuples;
/* estimate local relation sizes. */
tuples = get_local_rows(tuples,
path->path.multiple,
IsLocatorReplicated(path->path.locator_type),
ng_get_dest_num_data_nodes(&path->path)) /
dop;
} else
tuples = selec * outer_tuples * inner_tuples;
/* free space used by extended statistic */
if (es != NULL) {
qual_list = NIL;
list_free_ext(es->unmatched_clause_group);
delete es;
MemoryContextDelete(ExtendedStat);
}
ereport(DEBUG1,
(errmodule(MOD_OPT_JOIN),
errmsg("hashjointuples=%.0f, outer_tuples=%.0f, inner_tuples=%.0f, outer_global_tuples=%.0f, "
"inner_global_tuples=%.0f, selec=%.10f, multiple=%.0f,",
tuples,
outer_tuples,
inner_tuples,
outer_global_tuples,
inner_global_tuples,
selec,
path->path.multiple)));
return clamp_row_est(tuples);
}
/*
* set_baserel_size_estimates
* Set the size estimates for the given base relation.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already, and rel->tuples must be set.
*
* We set the following fields of the rel node:
* rows: the estimated number of output tuples (after applying
* restriction clauses).
* width: the estimated average output tuple width in bytes.
* baserestrictcost: estimated cost of evaluating baserestrictinfo clauses.
*/
void set_baserel_size_estimates(PlannerInfo* root, RelOptInfo* rel)
{
double nrows;
/* Should only be applied to base relations */
AssertEreport(
rel->relid > 0, MOD_OPT, "The relid is invalid when set the size estimates for the given base relation.");
nrows = rel->tuples * clauselist_selectivity(root, rel->baserestrictinfo, 0, JOIN_INNER, NULL);
rel->rows = clamp_row_est(nrows);
cost_qual_eval(&rel->baserestrictcost, rel->baserestrictinfo, root);
set_rel_width(root, rel);
}
/*
* get_parameterized_baserel_size
* Make a size estimate for a parameterized scan of a base relation.
*
* 'param_clauses' lists the additional join clauses to be used.
*
* set_baserel_size_estimates must have been applied already.
*/
double get_parameterized_baserel_size(PlannerInfo* root, RelOptInfo* rel, List* param_clauses)
{
List* allclauses = NIL;
double nrows;
/*
* Estimate the number of rows returned by the parameterized scan, knowing
* that it will apply all the extra join clauses as well as the rel's own
* restriction clauses. Note that we force the clauses to be treated as
* non-join clauses during selectivity estimation.
*/
allclauses = list_concat(list_copy(param_clauses), rel->baserestrictinfo);
nrows = rel->tuples * clauselist_selectivity(root,
allclauses,
rel->relid, /* do not use 0! */
JOIN_INNER,
NULL,
false);
nrows = clamp_row_est(nrows);
/* For safety, make sure result is not more than the base estimate */
if (nrows > rel->rows)
nrows = rel->rows;
return nrows;
}
/*
* set_joinrel_size_estimates
* Set the size estimates for the given join relation.
*
* The rel's targetlist must have been constructed already, and a
* restriction clause list that matches the given component rels must
* be provided.
*
* Since there is more than one way to make a joinrel for more than two
* base relations, the results we get here could depend on which component
* rel pair is provided. In theory we should get the same answers no matter
* which pair is provided; in practice, since the selectivity estimation
* routines don't handle all cases equally well, we might not. But there's
* not much to be done about it. (Would it make sense to repeat the
* calculations for each pair of input rels that's encountered, and somehow
* average the results? Probably way more trouble than it's worth, and
* anyway we must keep the rowcount estimate the same for all paths for the
* joinrel.)
*
* We set only the rows field here. The width field was already set by
* build_joinrel_tlist, and baserestrictcost is not used for join rels.
*/
void set_joinrel_size_estimates(PlannerInfo* root, RelOptInfo* rel, RelOptInfo* outer_rel, RelOptInfo* inner_rel,
SpecialJoinInfo* sjinfo, List* restrictlist)
{
rel->rows = calc_joinrel_size_estimate(root, outer_rel->rows, inner_rel->rows, sjinfo, restrictlist, true);
/*
* We should adjust joinrel's rows as max between outer_rel and inner_rel
* when the join rel's global rows is over estimate(more than 1.0e11).
*/
if (((unsigned int)u_sess->attr.attr_sql.cost_param & COST_ALTERNATIVE_CONJUNCT) &&
(rel->rows >= JOINREL_MAX_GLOBAL_ROWS) && (1 < list_length(restrictlist))) {
rel->rows = clamp_row_est(Min(Min(Max(outer_rel->rows, inner_rel->rows), rel->rows), JOINREL_MAX_GLOBAL_ROWS));
}
/* If global rows is less more, we should adjust it, unless either one is replicate */
if (IS_STREAM_PLAN && (rel->rows == 1) &&
!(IsLocatorReplicated(inner_rel->locator_type) || IsLocatorReplicated(outer_rel->locator_type))) {
Distribution* distribution = ng_get_default_computing_group_distribution();
rel->rows = bms_num_members(distribution->bms_data_nodeids);
}
rel->tuples = rel->rows;
}
/*
* set_joinpath_multiple_for_EC
* Set multiple for joinpath when the inner path or outer path contains
* EC function without following by StreamPath
*
* Because EC function only execute on one single datanode, so we should set
* the multiple particularly. When there is no StreamPath after the EC FunctionScan,
* the join result very likely be skew, we set joinpath multiple in this scene.
*/
void set_joinpath_multiple_for_EC(PlannerInfo* root, Path* path, Path* outer_path, Path* inner_path)
{
if (path == NULL || outer_path == NULL || inner_path == NULL || root == NULL) {
return;
}
RangeTblEntry* rte = NULL;
if (outer_path->pathtype == T_FunctionScan) {
rte = planner_rt_fetch(outer_path->parent->relid, root);
if (IS_EC_FUNC(rte)) {
path->multiple = outer_path->parent->multiple;
}
} else if (inner_path->pathtype == T_FunctionScan) {
rte = planner_rt_fetch(inner_path->parent->relid, root);
if (IS_EC_FUNC(rte)) {
path->multiple = inner_path->parent->multiple;
}
}
return;
}
/*
* get_parameterized_joinrel_size
* Make a size estimate for a parameterized scan of a join relation.
*
* 'rel' is the joinrel under consideration.
* 'outer_rows', 'inner_rows' are the sizes of the (probably also
* parameterized) join inputs under consideration.
* 'sjinfo' is any SpecialJoinInfo relevant to this join.
* 'restrict_clauses' lists the join clauses that need to be applied at the
* join node (including any movable clauses that were moved down to this join,
* and not including any movable clauses that were pushed down into the
* child paths).
*
* set_joinrel_size_estimates must have been applied already.
*/
double get_parameterized_joinrel_size(PlannerInfo* root, RelOptInfo* rel, double outer_rows, double inner_rows,
SpecialJoinInfo* sjinfo, List* restrict_clauses)
{
double nrows;
/*
* Estimate the number of rows returned by the parameterized join as the
* sizes of the input paths times the selectivity of the clauses that have
* ended up at this join node.
*
* As with set_joinrel_size_estimates, the rowcount estimate could depend
* on the pair of input paths provided, though ideally we'd get the same
* estimate for any pair with the same parameterization.
*/
nrows = calc_joinrel_size_estimate(root, outer_rows, inner_rows, sjinfo, restrict_clauses, false);
/* For safety, make sure result is not more than the base estimate */
if (nrows > rel->rows)
nrows = rel->rows;
return nrows;
}
/*
* calc_joinrel_size_estimate
* Workhorse for set_joinrel_size_estimates and
* get_parameterized_joinrel_size.
*/
static double calc_joinrel_size_estimate(PlannerInfo* root, double outer_rows, double inner_rows,
SpecialJoinInfo* sjinfo, List* restrictlist, bool varratio_cached)
{
JoinType jointype = sjinfo->jointype;
Selectivity jselec;
Selectivity pselec;
double nrows;
/*
* Compute joinclause selectivity. Note that we are only considering
* clauses that become restriction clauses at this join level; we are not
* double-counting them because they were not considered in estimating the
* sizes of the component rels.
*
* For an outer join, we have to distinguish the selectivity of the join's
* own clauses (JOIN/ON conditions) from any clauses that were "pushed
* down". For inner joins we just count them all as joinclauses.
*/
if (IS_OUTER_JOIN((uint32)jointype)) {
List* joinquals = NIL;
List* pushedquals = NIL;
ListCell* l = NULL;
/* Grovel through the clauses to separate into two lists */
foreach (l, restrictlist) {
RestrictInfo* rinfo = (RestrictInfo*)lfirst(l);
AssertEreport(IsA(rinfo, RestrictInfo),
MOD_OPT,
"The nodeTag of rinfo is T_RestrictInfo"
"when calculating the joinrel size.");
if (rinfo->is_pushed_down)
pushedquals = lappend(pushedquals, rinfo);
else
joinquals = lappend(joinquals, rinfo);
}
/* Get the separate selectivities */
jselec = clauselist_selectivity(root, joinquals, 0, jointype, sjinfo, varratio_cached);
pselec = clauselist_selectivity(root, pushedquals, 0, jointype, sjinfo, varratio_cached);
/* Avoid leaking a lot of ListCells */
list_free_ext(joinquals);
list_free_ext(pushedquals);
} else {
jselec = clauselist_selectivity(root, restrictlist, 0, jointype, sjinfo, varratio_cached);
pselec = 0.0; /* not used, keep compiler quiet */
}
/*
* Basically, we multiply size of Cartesian product by selectivity.
*
* If we are doing an outer join, take that into account: the joinqual
* selectivity has to be clamped using the knowledge that the output must
* be at least as large as the non-nullable input. However, any
* pushed-down quals are applied after the outer join, so their
* selectivity applies fully.
*
* For JOIN_SEMI and JOIN_ANTI, the selectivity is defined as the fraction
* of LHS rows that have matches, and we apply that straightforwardly.
*/
switch (jointype) {
case JOIN_INNER:
nrows = outer_rows * inner_rows * jselec;
break;
case JOIN_LEFT:
nrows = outer_rows * inner_rows * jselec;
if (nrows < outer_rows)
nrows = outer_rows;
nrows *= pselec;
break;
case JOIN_FULL:
nrows = outer_rows * inner_rows * jselec;
if (nrows < outer_rows)
nrows = outer_rows;
if (nrows < inner_rows)
nrows = inner_rows;
nrows *= pselec;
break;
case JOIN_SEMI:
nrows = outer_rows * jselec;
/* pselec not used */
break;
case JOIN_ANTI:
case JOIN_LEFT_ANTI_FULL:
nrows = outer_rows * (1.0 - jselec);
nrows *= pselec;
break;
default: {
/* other values not expected here */
ereport(ERROR,
(errmodule(MOD_OPT),
errcode(ERRCODE_UNRECOGNIZED_NODE_TYPE),
errmsg("unrecognized join type when calculate joinrel size estimate: %d", (int)jointype)));
nrows = 0; /* keep compiler quiet */
} break;
}
return clamp_row_est(nrows);
}
/*
* set_subquery_size_estimates
* Set the size estimates for a base relation that is a subquery.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already, and the plan for the subquery must have been completed.
* We look at the subquery's plan and PlannerInfo to extract data.
*
* We set the same fields as set_baserel_size_estimates.
*/
void set_subquery_size_estimates(PlannerInfo* root, RelOptInfo* rel)
{
PlannerInfo* subroot = rel->subroot;
RangeTblEntry PG_USED_FOR_ASSERTS_ONLY* rte = NULL;
ListCell* lc = NULL;
/* Should only be applied to base relations that are subqueries */
AssertEreport(rel->relid > 0,
MOD_OPT,
"The relid is invalid when set the size estimates for a base relation that is a subquery.");
rte = planner_rt_fetch(rel->relid, root);
AssertEreport(rte->rtekind == RTE_SUBQUERY,
MOD_OPT,
"Only subquery in FROM clause can be supported"
"when set the size estimates for a base relation that is a subquery.");
/* Copy raw number of output rows from subplan */
if (rel->subplan->exec_nodes != NULL)
rel->locator_type = rel->subplan->exec_nodes->baselocatortype;
if (rel->locator_type != LOCATOR_TYPE_REPLICATED)
rel->tuples = rel->subplan->plan_rows;
else
rel->tuples = PLAN_LOCAL_ROWS(rel->subplan);
set_local_rel_size(root, rel);
/*
* Compute per-output-column width estimates by examining the subquery's
* targetlist. For any output that is a plain Var, get the width estimate
* that was made while planning the subquery. Otherwise, we leave it to
* set_rel_width to fill in a datatype-based default estimate.
*/
foreach (lc, subroot->parse->targetList) {
TargetEntry* te = (TargetEntry*)lfirst(lc);
Node* texpr = (Node*)te->expr;
int32 item_width = 0;
AssertEreport(IsA(te, TargetEntry),
MOD_OPT,
"The nodeTag of te is T_TargetEntry"
"when set the size estimates for a base relation that is a subquery.");
/* junk columns aren't visible to upper query */
if (te->resjunk)
continue;
/*
* The subquery could be an expansion of a view that's had columns
* added to it since the current query was parsed, so that there are
* non-junk tlist columns in it that don't correspond to any column
* visible at our query level. Ignore such columns.
*/
if (te->resno < rel->min_attr || te->resno > rel->max_attr)
continue;
/*
* XXX This currently doesn't work for subqueries containing set
* operations, because the Vars in their tlists are bogus references
* to the first leaf subquery, which wouldn't give the right answer
* even if we could still get to its PlannerInfo.
*
* Also, the subquery could be an appendrel for which all branches are
* known empty due to constraint exclusion, in which case
* set_append_rel_pathlist will have left the attr_widths set to zero.
*
* In either case, we just leave the width estimate zero until
* set_rel_width fixes it.
*/
if (IsA(texpr, Var) && subroot->parse->setOperations == NULL) {
Var* var = (Var*)texpr;
RelOptInfo* subrel = find_base_rel(subroot, var->varno);
item_width = subrel->attr_widths[var->varattno - subrel->min_attr];
}
rel->attr_widths[te->resno - rel->min_attr] = item_width;
}
/* Now estimate number of output rows, etc */
set_baserel_size_estimates(root, rel);
}
/*
* set_function_size_estimates
* Set the size estimates for a base relation that is a function call.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already.
*
* We set the same fields as set_baserel_size_estimates.
*/
void set_function_size_estimates(PlannerInfo* root, RelOptInfo* rel)
{
RangeTblEntry* rte = NULL;
/* Should only be applied to base relations that are functions */
AssertEreport(rel->relid > 0,
MOD_OPT,
"The relid is invalid when set the size estimates for a base relation that is a function call.");
rte = planner_rt_fetch(rel->relid, root);
AssertEreport(rte->rtekind == RTE_FUNCTION,
MOD_OPT,
"Only function in FROM clause can be supported"
"when set the size estimates for a base relation that is a subquery.");
/* Estimate number of rows the function itself will return */
rel->tuples = expression_returns_set_rows(rte->funcexpr);
set_local_rel_size(root, rel);
/* Now estimate number of output rows, etc */
set_baserel_size_estimates(root, rel);
}
/*
* set_values_size_estimates
* Set the size estimates for a base relation that is a values list.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already.
*
* We set the same fields as set_baserel_size_estimates.
*/
void set_values_size_estimates(PlannerInfo* root, RelOptInfo* rel)
{
RangeTblEntry* rte = NULL;
/* Should only be applied to base relations that are values lists */
AssertEreport(rel->relid > 0,
MOD_OPT,
"The relid is invalid when set the size estimates for a base relation that is a values list.");
rte = planner_rt_fetch(rel->relid, root);
AssertEreport(rte->rtekind == RTE_VALUES,
MOD_OPT,
"Only VALUES list can be supported"
"when set the size estimates for a base relation that is a values list.");
/*
* Estimate number of rows the values list will return. We know this
* precisely based on the list length (well, barring set-returning
* functions in list items, but that's a refinement not catered for
* anywhere else either).
*/
rel->tuples = list_length(rte->values_lists);
set_local_rel_size(root, rel);
/* Now estimate number of output rows, etc */
set_baserel_size_estimates(root, rel);
}
/*
* set_cte_size_estimates
* Set the size estimates for a base relation that is a CTE reference.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already, and we need the completed plan for the CTE (if a regular CTE)
* or the non-recursive term (if a self-reference).
*
* We set the same fields as set_baserel_size_estimates.
*/
void set_cte_size_estimates(PlannerInfo* root, RelOptInfo* rel, Plan* cteplan)
{
RangeTblEntry* rte = NULL;
/* Should only be applied to base relations that are CTE references */
AssertEreport(rel->relid > 0,
MOD_OPT,
"The relid is invalid when set the size estimates for a base relation that is a CTE reference.");
rte = planner_rt_fetch(rel->relid, root);
AssertEreport(rte->rtekind == RTE_CTE,
MOD_OPT,
"Only common table expr can be supported"
"when set the size estimates for a base relation that is a CTE reference.");
if (rte->self_reference) {
/*
* In a self-reference, arbitrarily assume the average worktable size
* is about 10 times the nonrecursive term's size.
*/
rel->tuples = 10 * cteplan->plan_rows;
} else {
/* Otherwise just believe the CTE plan's output estimate */
rel->tuples = cteplan->plan_rows;
}
set_local_rel_size(root, rel);
/* Now estimate number of output rows, etc */
set_baserel_size_estimates(root, rel);
}
/*
* set_foreign_size_estimates
* Set the size estimates for a base relation that is a foreign table.
*
* There is not a whole lot that we can do here; the foreign-data wrapper
* is responsible for producing useful estimates. We can do a decent job
* of estimating baserestrictcost, so we set that, and we also set up width
* using what will be purely datatype-driven estimates from the targetlist.
* There is no way to do anything sane with the rows value, so we just put
* a default estimate and hope that the wrapper can improve on it. The
* wrapper's GetForeignRelSize function will be called momentarily.
*
* The rel's targetlist and restrictinfo list must have been constructed
* already.
*/
void set_foreign_size_estimates(PlannerInfo* root, RelOptInfo* rel)
{
/* Should only be applied to base relations */
AssertEreport(rel->relid > 0,
MOD_OPT,
"The relid is invalid when set the size estimates for a base relation that is a foreign table.");
rel->rows = 1000; /* entirely bogus default estimate */
cost_qual_eval(&rel->baserestrictcost, rel->baserestrictinfo, root);
set_rel_width(root, rel);
}
/*
* set_rel_width
* Set the estimated output width of a base relation.
*
* The estimated output width is the sum of the per-attribute width estimates
* for the actually-referenced columns, plus any PHVs or other expressions
* that have to be calculated at this relation. This is the amount of data
* we'd need to pass upwards in case of a sort, hash, etc.
*
* NB: this works best on plain relations because it prefers to look at
* real Vars. For subqueries, set_subquery_size_estimates will already have
* copied up whatever per-column estimates were made within the subquery,
* and for other types of rels there isn't much we can do anyway. We fall
* back on (fairly stupid) datatype-based width estimates if we can't get
* any better number.
*
* The per-attribute width estimates are cached for possible re-use while
* building join relations.
*/
void set_rel_width(PlannerInfo* root, RelOptInfo* rel)
{
Oid reloid = planner_rt_fetch(rel->relid, root)->relid;
int32 tuple_width = 0;
bool have_wholerow_var = false;
ListCell* lc = NULL;
foreach (lc, rel->reltargetlist) {
Node* node = (Node*)lfirst(lc);
if (IsA(node, Var)) {
Var* var = (Var*)node;
int ndx;
int32 item_width;
AssertEreport(var->varno == rel->relid,
MOD_OPT,
"The varno does not match to relid when setting the estimated output width of a base relation.");
AssertEreport(var->varattno >= rel->min_attr,
MOD_OPT,
"The varattno is less than min_attr when setting the estimated output width of a base relation.");
AssertEreport(var->varattno <= rel->max_attr,
MOD_OPT,
"The varattno is larger than max_attr when setting the estimated output width of a base relation.");
ndx = var->varattno - rel->min_attr;
/*
* If it's a whole-row Var, we'll deal with it below after we have
* already cached as many attr widths as possible.
*/
if (var->varattno == 0) {
have_wholerow_var = true;
continue;
}
/*
* The width may have been cached already (especially if it's a
* subquery), so don't duplicate effort.
*/
if (rel->attr_widths[ndx] > 0) {
tuple_width += rel->attr_widths[ndx];
if (root->glob->vectorized) {
rel->encodedwidth += columnar_get_col_width(var->vartype, rel->attr_widths[ndx]);
rel->encodednum++;
}
continue;
}
/* Try to get column width from statistics */
if (reloid != InvalidOid && var->varattno > 0) {
Oid targetid = reloid;
bool ispartition = false;
RangeTblEntry* rte = planner_rt_fetch(rel->relid, root);
if (rte->isContainPartition) {
AssertEreport(OidIsValid(rte->partitionOid),
MOD_OPT,
"The partitionOid is invalid when setting the estimated output width of a base relation.");
targetid = rte->partitionOid;
ispartition = true;
}
item_width = get_attavgwidth(targetid, var->varattno, ispartition);
if (item_width > 0) {
rel->attr_widths[ndx] = item_width;
tuple_width += item_width;
if (root->glob->vectorized) {
rel->encodedwidth += columnar_get_col_width(var->vartype, item_width);
rel->encodednum++;
}
continue;
}
}
/*
* Not a plain relation, or can't find statistics for it. Estimate
* using just the type info.
*/
item_width = get_typavgwidth(var->vartype, var->vartypmod);
AssertEreport(item_width > 0,
MOD_OPT,
"The estimated average width of values of the type is not larger than 0"
"when setting the estimated output width of a base relation.");
rel->attr_widths[ndx] = item_width;
tuple_width += item_width;
if (root->glob->vectorized) {
rel->encodedwidth += columnar_get_col_width(var->vartype, item_width);
rel->encodednum++;
}
} else if (IsA(node, PlaceHolderVar)) {
PlaceHolderVar* phv = (PlaceHolderVar*)node;
PlaceHolderInfo* phinfo = find_placeholder_info(root, phv, false);
tuple_width += phinfo->ph_width;
if (root->glob->vectorized) {
rel->encodedwidth += columnar_get_col_width(exprType((Node*)phv->phexpr), phinfo->ph_width);
rel->encodednum++;
}
} else {
/*
* We could be looking at an expression pulled up from a subquery,
* or a ROW() representing a whole-row child Var, etc. Do what we
* can using the expression type information.
*/
int32 item_width;
item_width = get_typavgwidth(exprType(node), exprTypmod(node));
AssertEreport(item_width > 0,
MOD_OPT,
"The estimated average width of values of the type is not larger than 0"
"when setting the estimated output width of a base relation.");
tuple_width += item_width;
if (root->glob->vectorized) {
rel->encodedwidth += columnar_get_col_width(exprType(node), item_width);
rel->encodednum++;
}
}
}
/*
* If we have a whole-row reference, estimate its width as the sum of
* per-column widths plus sizeof(HeapTupleHeaderData).
*/
if (have_wholerow_var) {
int32 wholerow_width = sizeof(HeapTupleHeaderData);
if (reloid != InvalidOid) {
Oid partid = InvalidOid;
RangeTblEntry* rte = planner_rt_fetch(rel->relid, root);
if (rte->isContainPartition) {
AssertEreport(OidIsValid(rte->partitionOid),
MOD_OPT,
"The partitionOid is invalid when setting the estimated output width of a base relation.");
partid = rte->partitionOid;
}
/* Real relation, so estimate true tuple width */
wholerow_width += get_relation_data_width(reloid, partid, rel->attr_widths - rel->min_attr);
} else {
/* Do what we can with info for a phony rel */
AttrNumber i;
for (i = 1; i <= rel->max_attr; i++)
wholerow_width += rel->attr_widths[i - rel->min_attr];
}
rel->attr_widths[0 - rel->min_attr] = wholerow_width;
/*
* Include the whole-row Var as part of the output tuple. Yes, that
* really is what happens at runtime.
*/
tuple_width += wholerow_width;
}
AssertEreport(tuple_width >= 0,
MOD_OPT,
"The estimated width of tuple is less than 0"
"when setting the estimated output width of a base relation.");
rel->width = tuple_width;
}
/*
* relation_byte_size
* Estimate the storage space in bytes for a given number of tuples
* of a given width (size in bytes).
*
* Parameters:
* @in tuples: number of rows of input relation
* @in width: width of input relation
* @in vectorized: if rel is vectorized stored
* @in aligned: should be calculated as stored aligned. For disk storage,
* it's false, but in execution, the memory should be aligned
* @in issort: if the path is sort. Since the function will be used by
* sort and materialize, and their way to calculate width is different
* @in index_sort: if index header should be used instead of heap
*
* Returns: total size of the input relation
*/
double relation_byte_size(double tuples, int width, bool vectorized, bool aligned, bool issort, bool indexsort)
{
Assert(width >= 0);
size_t header_size = (issort && indexsort) ? sizeof(IndexTupleData) : sizeof(HeapTupleHeaderData);
if (aligned) {
if (vectorized)
return tuples * (TUPLE_OVERHEAD(true) + width);
else
return tuples *
(TUPLE_OVERHEAD(issort) + alloc_trunk_size(MAXALIGN((uintptr_t)width) + MAXALIGN(header_size)));
} else {
return tuples * (MAXALIGN((uintptr_t)width) + MAXALIGN(header_size));
}
}
/*
* page_size
* Returns an estimate of the number of pages covered by a given
* number of tuples of a given width (size in bytes).
*/
double page_size(double tuples, int width)
{
return ceil(relation_byte_size(tuples, width, false) / BLCKSZ);
}
/*
* Estimate the fraction of the work that each worker will do given the
* number of workers budgeted for the path.
*/
static double get_parallel_divisor(Path* path)
{
double parallel_divisor = path->parallel_workers;
/*
* Early experience with parallel query suggests that when there is only
* one worker, the leader often makes a very substantial contribution to
* executing the parallel portion of the plan, but as more workers are
* added, it does less and less, because it's busy reading tuples from the
* workers and doing whatever non-parallel post-processing is needed. By
* the time we reach 4 workers, the leader no longer makes a meaningful
* contribution. Thus, for now, estimate that the leader spends 30% of
* its time servicing each worker, and the remainder executing the
* parallel plan.
*/
if (u_sess->attr.attr_sql.parallel_leader_participation) {
double leader_contribution;
leader_contribution = 1.0 - (0.3 * path->parallel_workers);
if (leader_contribution > 0) {
parallel_divisor += leader_contribution;
}
}
return parallel_divisor;
}
/*
* compute_bitmap_pages
*
* compute number of pages fetched from heap in bitmap heap scan.
*/
double compute_bitmap_pages(PlannerInfo *root, RelOptInfo *baserel, Path *bitmapqual,
double loop_count, Cost *cost, double *tuple, bool ispartitionedindex)
{
Cost indexTotalCost;
Selectivity indexSelectivity;
double pages_fetched;
double T = (baserel->pages > 1) ? (double)baserel->pages : 1.0;
/*
* Fetch total cost of obtaining the bitmap, as well as its total
* selectivity.
*/
cost_bitmap_tree_node(bitmapqual, &indexTotalCost, &indexSelectivity);
/*
* Estimate number of main-table pages fetched.
*/
double tuples_fetched = clamp_row_est(indexSelectivity * RELOPTINFO_LOCAL_FIELD(root, baserel, tuples));
if (loop_count > 1) {
/*
* For repeated bitmap scans, scale up the number of tuples fetched in
* the Mackert and Lohman formula by the number of scans, so that we
* estimate the number of pages fetched by all the scans. Then
* pro-rate for one scan.
*/
pages_fetched = index_pages_fetched(tuples_fetched * loop_count,
(BlockNumber)baserel->pages,
get_indexpath_pages(bitmapqual),
root,
ispartitionedindex);
pages_fetched /= loop_count;
} else {
/*
* For a single scan, the number of heap pages that need to be fetched
* is the same as the Mackert and Lohman formula for the case T <= b
* (ie, no re-reads needed).
*/
pages_fetched = (2.0 * T * tuples_fetched) / (2.0 * T + tuples_fetched);
}
if (pages_fetched >= T) {
pages_fetched = T;
} else {
pages_fetched = ceil(pages_fetched);
}
if (cost != NULL) {
*cost = indexTotalCost;
}
if (tuple != NULL) {
*tuple = tuples_fetched;
}
return pages_fetched;
}
/* it used to compute page_size in createplan.cpp */
double cost_page_size(double tuples, int width)
{
return page_size(tuples, width);
}
/*
* restore_hashjoin_cost
* With u_sess->attr.attr_sql.enable_change_hjcost on, we lessen startup cost to do the internal cost comparison.
* Before final judgement, we should restore it back.
*/
void restore_hashjoin_cost(Path* path)
{
if (u_sess->attr.attr_sql.enable_change_hjcost && IsA(path, HashPath)) {
Path* innerpath = ((HashPath*)path)->jpath.innerjoinpath;
path->startup_cost += innerpath->total_cost - innerpath->startup_cost;
}
}
/*
* finalize_dml_cost:
* calculate insert/update/delete cost of the modifytable plan
* Param:
* @in plan: target modifytable plan
*/
void finalize_dml_cost(ModifyTable* plan)
{
CmdType type = plan->operation;
/* For insert and update, we should insert the tuples, so add the insertion cost */
if (type == CMD_INSERT || type == CMD_UPDATE) {
double size = page_size(PLAN_LOCAL_ROWS(&plan->plan), plan->plan.plan_width);
plan->plan.total_cost +=
(u_sess->attr.attr_sql.cpu_tuple_cost + u_sess->attr.attr_sql.seq_page_cost * 2) * size;
}
/* For delete and update we should scan the tuples by ctid, so add random page cost */
if (type == CMD_DELETE || type == CMD_UPDATE) {
plan->plan.total_cost += u_sess->attr.attr_sql.random_page_cost * PLAN_LOCAL_ROWS(&plan->plan);
}
}
/*
* Description: Get sample fraction.
*
* Parameters:
* @in pctnode: node of percent args.
*
* Return: float4
*/
static float4 get_samplefract(Node* pctnode)
{
float4 samplefract;
if (IsA(pctnode, Const) && !((Const*)pctnode)->constisnull) {
samplefract = DatumGetFloat4(((Const*)pctnode)->constvalue);
if (samplefract >= 0.0 && samplefract <= 100.0 && !isnan(samplefract)) {
samplefract /= 100.0f;
} else {
/* Default samplefract if the value is bogus */
samplefract = 0.1f;
}
} else {
/* Default samplefract if we didn't obtain a non-null Const */
samplefract = 0.1f;
}
return samplefract;
}
/*
* Description: Sample size estimation.
*
* Parameters:
* @in root: plannerinfo struct for current query level.
* @in baserel: the relation to be scanned.
* @in paramexprs: the sample percentage info.
*
* Return: void
*/
void system_samplescangetsamplesize(PlannerInfo* root, RelOptInfo* baserel, List* paramexprs)
{
Node* pctnode = NULL;
float4 samplefract;
/* Try to extract an estimate for the sample percentage */
pctnode = (Node*)linitial(paramexprs);
pctnode = estimate_expression_value(root, pctnode);
samplefract = get_samplefract(pctnode);
/* We'll visit a sample of the pages ... */
baserel->pages = clamp_row_est(baserel->pages * samplefract);
/* ... and hopefully get a representative number of tuples from them */
baserel->tuples = clamp_row_est(baserel->tuples * samplefract);
}
/*
* Description: Sample size estimation.
*
* Parameters:
* @in root: plannerinfo struct for current query level.
* @in baserel: the relation to be scanned.
* @in paramexprs: the sample percentage info.
*
* Return: void
*/
void bernoulli_samplescangetsamplesize(PlannerInfo* root, RelOptInfo* baserel, List* paramexprs)
{
Node* pctnode = NULL;
float4 samplefract;
/* Try to extract an estimate for the sample percentage */
pctnode = (Node*)linitial(paramexprs);
pctnode = estimate_expression_value(root, pctnode);
samplefract = get_samplefract(pctnode);
/* We'll visit all pages of the baserel, so pages is the same */
baserel->tuples = clamp_row_est(baserel->tuples * samplefract);
}
/*
* Description: Sample size estimation.
*
* Parameters:
* @in root: plannerinfo struct for current query level.
* @in baserel: the relation to be scanned.
* @in paramexprs: the sample percentage info.
*
* Return: void
*/
void hybrid_samplescangetsamplesize(PlannerInfo* root, RelOptInfo* baserel, List* paramexprs)
{
ListCell* lc = NULL;
uint16 i = 0;
AssertEreport(SAMPLEARGSNUM == list_length(paramexprs),
MOD_OPT,
"The number of sample percentage info does not equal 2"
"when setting the estimated output width of a base relation.");
foreach (lc, paramexprs) {
Node* paramnode = (Node*)lfirst(lc);
Node* pctnode = estimate_expression_value(root, paramnode);
float4 samplefract = get_samplefract(pctnode);
if (i == SYSTEM_SAMPLE) {
/* We'll visit a sample of the pages ... */
baserel->pages = clamp_row_est(baserel->pages * samplefract);
}
/* ... and hopefully get a representative number of tuples from them */
baserel->tuples = clamp_row_est(baserel->tuples * samplefract);
i++;
}
}
/*
* copy_mem_info
* copy OpMemInfo structure from source to dest
*
* Parameters:
* @in dest: dest structure
* @in src: source structure
*
* Returns: void
*/
void copy_mem_info(OpMemInfo* dest, OpMemInfo* src)
{
errno_t rc = 0;
rc = memcpy_s(dest, sizeof(OpMemInfo), src, sizeof(OpMemInfo));
securec_check(rc, "\0", "\0");
}
/*
* columnar_get_col_width
* Calculate additional space besides datum for a columnar column,
* only happens on length-varying columns
*
* Parameters:
* @in typid: column type oid
* @in width: column raw width
* @in aligned: if return value is aligned by memory aset module
*
* Returns: additional width for the column, or 0 for fixed-length col
*/
int columnar_get_col_width(Oid typid, int width, bool aligned)
{
if (COL_IS_ENCODE(typid)) {
if (aligned) {
return alloc_trunk_size(width);
} else {
return width;
}
} else
return 0;
}
/*
* has_complicate_hashkey
* Judge if has complicate hash key, if so, vector engine will use another
* 8 bytes to store hash value to avoid duplicate calculation. Only simple
* var is not complicate hash key
*
* Parameters:
* @in hashclauses: equal clauses of hash join
* @in inner_relids: relids of hashjoin inner table
*
* Returns: If there's complicate hash key in the hash condition
*/
bool has_complicate_hashkey(List* hashclauses, Relids inner_relids)
{
ListCell* lc = NULL;
foreach (lc, hashclauses) {
RestrictInfo* restrictinfo = (RestrictInfo*)lfirst(lc);
Node* innerkey = NULL;
AssertEreport(IsA(restrictinfo, RestrictInfo),
MOD_OPT,
"The nodeTag of restrictinfo is not T_RestrictInfo"
"when setting the estimated output width of a base relation.");
/*
* First we have to figure out which side of the hashjoin clause
* is the inner side.
*/
if (bms_is_subset(restrictinfo->right_relids, inner_relids)) {
innerkey = get_rightop(restrictinfo->clause);
} else {
AssertEreport(bms_is_subset(restrictinfo->left_relids, inner_relids),
MOD_OPT,
"The left relids is not subset of the relids of inner side"
"when setting the estimated output width of a base relation.");
innerkey = get_leftop(restrictinfo->clause);
}
/* Judge if the inner key is simple var */
if (innerkey != NULL && !IsA(innerkey, Var) &&
!(IsA(innerkey, RelabelType) && IsA(((RelabelType*)innerkey)->arg, Var)))
return true;
}
return false;
}
/*
* calc_distributekey_width
* Optimizer will add distribute key in the targetlist if not found in plan
* phase, so added width should be considerred here
*
* Parameters:
* @in path: the path that should estimate width
* @out width: return total width of added distribute keys
* @in vectorized: if the path will be vectorized plan
* @in aligned: if the width should be aligned by memory aset module
*
* Returns: number of distribute keys added to targetlist
*/
static int calc_distributekey_width(Path* path, int* width, bool vectorized, bool aligned)
{
int num = 0;
ListCell* lc = NULL;
/* Only do this for redistribute stream, since only redistribute has distribute keys */
if (IsA(path, StreamPath) && ((StreamPath*)path)->type == STREAM_REDISTRIBUTE) {
foreach (lc, path->distribute_keys) {
Node* node = (Node*)lfirst(lc);
if (!list_member(path->parent->reltargetlist, node)) {
num++;
int32 item_width = get_typavgwidth(exprType(node), exprTypmod(node));
AssertEreport(item_width > 0,
MOD_OPT,
"The item width is not larger than 0 when setting the estimated output width of a base relation.");
if (vectorized)
*width += columnar_get_col_width(exprType(node), item_width, aligned);
else
*width += item_width;
}
}
}
return num;
}
/*
* get_path_actual_total_width
* In PG optimizer, only width of row engine is estimated, and it has
* big difference with vector engine, so this function is used to estimate
* width of a path with vector engine though row width
*
* Parameters:
* @in path: the path that should estimate width
* @in vectorized: if the path will be vectorized
* @in type: the type to calculate the width, only considerring hashjoin,
* hashagg, sort and material
* @in newcol: for some case like add distribute column, and vectorized
* abbreviate sort, new columns will be added, so should record
* the number of new col to impact the width
*
* Returns: estimated width with row-engine and vector-engine
*/
int get_path_actual_total_width(Path* path, bool vectorized, OpType type, int newcol)
{
int num_new_col = 0;
int width = 0;
bool aligned = (type >= OP_SORT);
if (path->parent == NULL) {
return COL_TUPLE_WIDTH;
}
/* For redistribute, we will add unmatched distribute key into targetlist, so count this */
num_new_col = calc_distributekey_width(path, &width, vectorized, aligned);
if (vectorized) {
switch (type) {
case OP_HASHJOIN:
width += path->parent->encodedwidth +
SIZE_COL_VALUE * (list_length(path->parent->reltargetlist) + num_new_col + newcol);
break;
case OP_HASHAGG:
width += path->parent->encodedwidth + TUPLE_OVERHEAD(true) + sizeof(void*) * 2 +
SIZE_COL_VALUE * (list_length(path->parent->reltargetlist) + num_new_col + newcol);
break;
case OP_SORT:
if (width != 0 || path->parent->encodednum != 0)
newcol += 1;
/* No need break here. */
case OP_MATERIAL:
/* don't know encoded width of each column, just average them for a rough estimation */
if (path->parent->encodednum > 0)
width += path->parent->encodednum *
alloc_trunk_size(path->parent->encodedwidth / path->parent->encodednum);
width += sizeof(Datum) * (list_length(path->parent->reltargetlist) + num_new_col + newcol);
break;
default:
break;
}
} else {
width += path->parent->width;
}
return width;
}
/*
* get_subqueryscan_stream_cost
* get stream_cost of a subquery
*/
static Cost get_subqueryscan_stream_cost(Plan* subplan)
{
Cost stream_cost = 0;
if (subplan == NULL)
return stream_cost;
switch (nodeTag(subplan)) {
case T_HashJoin:
case T_VecHashJoin:
stream_cost = get_subqueryscan_stream_cost(subplan->lefttree);
break;
case T_NestLoop:
case T_VecNestLoop:
stream_cost = get_subqueryscan_stream_cost(subplan->righttree);
break;
case T_MergeJoin:
case T_VecMergeJoin:
stream_cost = get_subqueryscan_stream_cost(subplan->righttree);
break;
case T_Stream:
case T_VecStream:
stream_cost = subplan->startup_cost;
break;
default:
stream_cost = get_subqueryscan_stream_cost(subplan->lefttree);
break;
}
return stream_cost;
}
Reference in New Issue View Git Blame Copy Permalink