Use imporved version of Postgres' simplehash table for aggregation #8006
Conversation
akuzm
commented
Apr 24, 2025
akuzm
commented
- No separate status field for better memory efficiency
- Slow path in insert function rewritten to use tail recursion for better inlining of fast path
* No separate status field for better memory efficiency * Slow path in insert function rewritten to use tail recursion for better inlining of fast path
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #8006 +/- ##
==========================================
- Coverage 82.21% 81.98% -0.24%
==========================================
Files 256 257 +1
Lines 48184 48303 +119
Branches 12155 12182 +27
==========================================
- Hits 39615 39599 -16
- Misses 3704 3891 +187
+ Partials 4865 4813 -52 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
Why do we need to vendor Postgres simplehash.h? |
We use it for vectorized aggregation, and this is a version where I made a couple peformance-related changes as described in the PR message. |
Ahh got it ... I've just quickly looked the both code and it looked like the same but actually no (after diff them). |
| /* | ||
| * This file and its contents are licensed under the Timescale License. | ||
| * Please see the included NOTICE for copyright information and | ||
| * LICENSE-TIMESCALE for a copy of the license. | ||
| */ |
There was a problem hiding this comment.
I know that Postgres license is very permissive, but doesn't make sense to vendor this code on Apache code?
| * PostgreSQL License. Please see the NOTICE at the top level | ||
| * directory for a copy of the PostgreSQL License. | ||
| * | ||
| * PostgreSQL 17.4 f8554dee417ffc4540c94cf357f7bf7d4b6e5d80 |
There was a problem hiding this comment.
Would be nice to have a brief explanation here about the changes you've made just for reference.
|
|
||
| /* <element> *<prefix>_lookup(<prefix>_hash *tb, <key> key) */ | ||
| SH_SCOPE SH_ELEMENT_TYPE *SH_LOOKUP(SH_TYPE * tb, SH_KEY_TYPE key); | ||
|
|
There was a problem hiding this comment.
Noticed that you removed SH_DELETE_ITEM and SH_DELETE. We don't need it? I'm just thinking that maybe this simplehash stuff can be used in other places.
