!2872 优化hashjoin的hash策略

Merge pull request !2872 from wanghao19920907/PR_2296_new
This commit is contained in:
opengauss-bot
2023-03-02 01:52:58 +00:00
committed by Gitee

View File

@ -37,6 +37,7 @@
#include "pgstat.h"
#include "pgxc/pgxc.h"
#include "instruments/instr_unique_sql.h"
#include "port/pg_bitutils.h"
#include "utils/anls_opt.h"
#include "utils/dynahash.h"
#include "utils/lsyscache.h"
@ -1424,7 +1425,7 @@ bool ExecHashGetHashValue(HashJoinTable hashtable, ExprContext* econtext, List*
* chains), and must only cause the batch number to remain the same or
* increase. Our algorithm is
* bucketno = hashvalue MOD nbuckets
* batchno = (hashvalue DIV nbuckets) MOD nbatch
* batchno = ROR(hashvalue, log2_nbuckets) MOD nbatch
* where nbuckets and nbatch are both expected to be powers of 2, so we can
* do the computations by shifting and masking. (This assumes that all hash
* functions are good about randomizing all their output bits, else we are
@ -1436,7 +1437,11 @@ bool ExecHashGetHashValue(HashJoinTable hashtable, ExprContext* econtext, List*
* number the way we do here).
*
* nbatch is always a power of 2; we increase it only by doubling it. This
* effectively adds one more bit to the top of the batchno.
* effectively adds one more bit to the top of the batchno. In very large
* joins, we might run out of bits to add, so we do this by rotating the hash
* value. This causes batchno to steal bits from bucketno when the number of
* virtual buckets exceeds 2^32. It's better to have longer bucket chains
* than to lose the ability to divide batches.
*/
void ExecHashGetBucketAndBatch(HashJoinTable hashtable, uint32 hashvalue, int* bucketno, int* batchno)
{
@ -1446,7 +1451,7 @@ void ExecHashGetBucketAndBatch(HashJoinTable hashtable, uint32 hashvalue, int* b
if (nbatch > 1) {
/* we can do MOD by masking, DIV by shifting */
*bucketno = hashvalue & (nbuckets - 1);
*batchno = (hashvalue >> hashtable->log2_nbuckets) & (nbatch - 1);
*batchno = pg_rotate_right32(hashvalue, hashtable->log2_nbuckets) & (nbatch - 1);
} else {
*bucketno = hashvalue & (nbuckets - 1);
*batchno = 0;