Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
colexec: use separate maps for each hash bucket in distinct hash agg
This commit changes the approach to how we verify whether a tuple is distinct. Previously, there was a single map for each aggregate function with DISTINCT clause that was shared among all groups. This required that we encode both grouping and aggregation columns of a tuple. Now each group (i.e. bucket) has a separate map for each distinct aggregate function. Initially I thought that this would incur too much of an overhead, but as it turns out, the "global" shared map grows a lot bigger and actually worsens the performance. Release note: None
- Loading branch information