Files
postgresql/src/backend/executor/nodeUnique.c
Tom Lane f97aebd162 Revise TupleTableSlot code to avoid unnecessary construction and disassembly
of tuples when passing data up through multiple plan nodes.  A slot can now
hold either a normal "physical" HeapTuple, or a "virtual" tuple consisting
of Datum/isnull arrays.  Upper plan levels can usually just copy the Datum
arrays, avoiding heap_formtuple() and possible subsequent nocachegetattr()
calls to extract the data again.  This work extends Atsushi Ogawa's earlier
patch, which provided the key idea of adding Datum arrays to TupleTableSlots.
(I believe however that something like this was foreseen way back in Berkeley
days --- see the old comment on ExecProject.)  A test case involving many
levels of join of fairly wide tables (about 80 columns altogether) showed
about 3x overall speedup, though simple queries will probably not be
helped very much.

I have also duplicated some code in heaptuple.c in order to provide versions
of heap_formtuple and friends that use "bool" arrays to indicate null
attributes, instead of the old convention of "char" arrays containing either
'n' or ' '.  This provides a better match to the convention used by
ExecEvalExpr.  While I have not made a concerted effort to get rid of uses
of the old routines, I think they should be deprecated and eventually removed.
2005-03-16 21:38:10 +00:00

206 lines
5.3 KiB
C

/*-------------------------------------------------------------------------
*
* nodeUnique.c
* Routines to handle unique'ing of queries where appropriate
*
* Portions Copyright (c) 1996-2005, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/executor/nodeUnique.c,v 1.46 2005/03/16 21:38:08 tgl Exp $
*
*-------------------------------------------------------------------------
*/
/*
* INTERFACE ROUTINES
* ExecUnique - generate a unique'd temporary relation
* ExecInitUnique - initialize node and subnodes..
* ExecEndUnique - shutdown node and subnodes
*
* NOTES
* Assumes tuples returned from subplan arrive in
* sorted order.
*/
#include "postgres.h"
#include "access/heapam.h"
#include "executor/executor.h"
#include "executor/nodeUnique.h"
/* ----------------------------------------------------------------
* ExecUnique
*
* This is a very simple node which filters out duplicate
* tuples from a stream of sorted tuples from a subplan.
* ----------------------------------------------------------------
*/
TupleTableSlot * /* return: a tuple or NULL */
ExecUnique(UniqueState *node)
{
Unique *plannode = (Unique *) node->ps.plan;
TupleTableSlot *resultTupleSlot;
TupleTableSlot *slot;
PlanState *outerPlan;
/*
* get information from the node
*/
outerPlan = outerPlanState(node);
resultTupleSlot = node->ps.ps_ResultTupleSlot;
/*
* now loop, returning only non-duplicate tuples. We assume that the
* tuples arrive in sorted order so we can detect duplicates easily.
*
* We return the first tuple from each group of duplicates (or the last
* tuple of each group, when moving backwards). At either end of the
* subplan, clear the result slot so that we correctly return the
* first/last tuple when reversing direction.
*/
for (;;)
{
/*
* fetch a tuple from the outer subplan
*/
slot = ExecProcNode(outerPlan);
if (TupIsNull(slot))
{
/* end of subplan; reset in case we change direction */
ExecClearTuple(resultTupleSlot);
return NULL;
}
/*
* Always return the first/last tuple from the subplan.
*/
if (TupIsNull(resultTupleSlot))
break;
/*
* Else test if the new tuple and the previously returned tuple
* match. If so then we loop back and fetch another new tuple
* from the subplan.
*/
if (!execTuplesMatch(slot, resultTupleSlot,
plannode->numCols, plannode->uniqColIdx,
node->eqfunctions,
node->tempContext))
break;
}
/*
* We have a new tuple different from the previous saved tuple (if
* any). Save it and return it. We must copy it because the source
* subplan won't guarantee that this source tuple is still accessible
* after fetching the next source tuple.
*/
return ExecCopySlot(resultTupleSlot, slot);
}
/* ----------------------------------------------------------------
* ExecInitUnique
*
* This initializes the unique node state structures and
* the node's subplan.
* ----------------------------------------------------------------
*/
UniqueState *
ExecInitUnique(Unique *node, EState *estate)
{
UniqueState *uniquestate;
/*
* create state structure
*/
uniquestate = makeNode(UniqueState);
uniquestate->ps.plan = (Plan *) node;
uniquestate->ps.state = estate;
/*
* Miscellaneous initialization
*
* Unique nodes have no ExprContext initialization because they never
* call ExecQual or ExecProject. But they do need a per-tuple memory
* context anyway for calling execTuplesMatch.
*/
uniquestate->tempContext =
AllocSetContextCreate(CurrentMemoryContext,
"Unique",
ALLOCSET_DEFAULT_MINSIZE,
ALLOCSET_DEFAULT_INITSIZE,
ALLOCSET_DEFAULT_MAXSIZE);
#define UNIQUE_NSLOTS 1
/*
* Tuple table initialization
*/
ExecInitResultTupleSlot(estate, &uniquestate->ps);
/*
* then initialize outer plan
*/
outerPlanState(uniquestate) = ExecInitNode(outerPlan(node), estate);
/*
* unique nodes do no projections, so initialize projection info for
* this node appropriately
*/
ExecAssignResultTypeFromOuterPlan(&uniquestate->ps);
uniquestate->ps.ps_ProjInfo = NULL;
/*
* Precompute fmgr lookup data for inner loop
*/
uniquestate->eqfunctions =
execTuplesMatchPrepare(ExecGetResultType(&uniquestate->ps),
node->numCols,
node->uniqColIdx);
return uniquestate;
}
int
ExecCountSlotsUnique(Unique *node)
{
return ExecCountSlotsNode(outerPlan(node)) +
ExecCountSlotsNode(innerPlan(node)) +
UNIQUE_NSLOTS;
}
/* ----------------------------------------------------------------
* ExecEndUnique
*
* This shuts down the subplan and frees resources allocated
* to this node.
* ----------------------------------------------------------------
*/
void
ExecEndUnique(UniqueState *node)
{
/* clean up tuple table */
ExecClearTuple(node->ps.ps_ResultTupleSlot);
MemoryContextDelete(node->tempContext);
ExecEndNode(outerPlanState(node));
}
void
ExecReScanUnique(UniqueState *node, ExprContext *exprCtxt)
{
/* must clear result tuple so first input tuple is returned */
ExecClearTuple(node->ps.ps_ResultTupleSlot);
/*
* if chgParam of subnode is not null then plan will be re-scanned by
* first ExecProcNode.
*/
if (((PlanState *) node)->lefttree->chgParam == NULL)
ExecReScan(((PlanState *) node)->lefttree, exprCtxt);
}