diff options
| author | Jeff Davis <jdavis@postgresql.org> | 2020-07-28 23:15:47 -0700 |
|---|---|---|
| committer | Jeff Davis <jdavis@postgresql.org> | 2020-07-28 23:16:28 -0700 |
| commit | 9878b643f37b1e4167f64a9941244bfabed60623 (patch) | |
| tree | f211bc0c45f888f8fea8d3827a2e2ae6be4fa0ea /src/include/executor/nodeAgg.h | |
| parent | f2130e77da51f35d37fd15a343bc1c4a4527e0fd (diff) | |
| download | postgresql-9878b643f37b1e4167f64a9941244bfabed60623.tar.gz postgresql-9878b643f37b1e4167f64a9941244bfabed60623.zip | |
HashAgg: use better cardinality estimate for recursive spilling.
Use HyperLogLog to estimate the group cardinality in a spilled
partition. This estimate is used to choose the number of partitions if
we recurse.
The previous behavior was to use the number of tuples in a spilled
partition as the estimate for the number of groups, which lead to
overpartitioning. That could cause the number of batches to be much
higher than expected (with each batch being very small), which made it
harder to interpret EXPLAIN ANALYZE results.
Reviewed-by: Peter Geoghegan
Discussion: https://postgr.es/m/a856635f9284bc36f7a77d02f47bbb6aaf7b59b3.camel@j-davis.com
Backpatch-through: 13
Diffstat (limited to 'src/include/executor/nodeAgg.h')
| -rw-r--r-- | src/include/executor/nodeAgg.h | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/include/executor/nodeAgg.h b/src/include/executor/nodeAgg.h index bb0805abe09..b9551695384 100644 --- a/src/include/executor/nodeAgg.h +++ b/src/include/executor/nodeAgg.h @@ -320,7 +320,7 @@ extern void ExecReScanAgg(AggState *node); extern Size hash_agg_entry_size(int numTrans, Size tupleWidth, Size transitionSpace); -extern void hash_agg_set_limits(double hashentrysize, uint64 input_groups, +extern void hash_agg_set_limits(double hashentrysize, double input_groups, int used_bits, Size *mem_limit, uint64 *ngroups_limit, int *num_partitions); |