feature: asynchronously update file statistics during vacuum #1397
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
gongxun0928
force-pushed
the
pax/collect-columns-stats-in-bgworker
branch
3 times, most recently
from
aaaecde to
382bc9a
Compare
gfphoenix78
reviewed
Oct 29, 2025
gongxun0928
force-pushed
the
pax/collect-columns-stats-in-bgworker
branch
5 times, most recently
from
9b508a6 to
1237fd6
Compare
Previously, statistics (min-max, sum, count, etc.) were computed synchronously during data insertion, causing significant slowdowns due to heavy computational overhead. This change introduces an asynchronous approach to maintain statistics: - Add a GUC parameter to control statistics collection during writes (disabled by default) - Skip statistics computation during INSERT to ensure fast writes - Update statistics asynchronously during VACUUM on PAX tables by scanning file metadata - Re-read files and refresh statistics only when metadata indicates they are stale - Vacuum which data files have been marked for deletion ``` create table t1(c1 int, c2 int, c3 int, c4 int, c5 int, c6 int) using pax with(minmax_columns='c1,c2,c3,c4,c5,c6'); set pax.enable_sync_collect_stats to on; -- collect stats synchronously insert into t1 select i,i,i,i,i,i from generate_series(1,1000000) i; INSERT 0 1000000 Time: 2733.731 ms (00:02.734) create table t2(c1 int, c2 int, c3 int, c4 int, c5 int, c6 int) using pax; insert into t2 select i,i,i,i,i,i from generate_series(1,1000000) i; INSERT 0 1000000 Time: 1816.836 ms (00:01.817) ```
gongxun0928
force-pushed
the
pax/collect-columns-stats-in-bgworker
branch
from
1237fd6 to
cd81591
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously, statistics (min-max, sum, count, etc.) were computed synchronously during data insertion, causing significant slowdowns due to heavy computational overhead.
This change introduces an asynchronous approach to maintain statistics:
In this PR, you'll notice changes in many test results related to VACUUM. The main reason is that after data has been updated or deleted, when VACUUM is performed, files marked for deletion will be rewritten—potentially consolidating multiple such files in the process.
It's important to note that this PR does not include the implementation of VACUUM FULL.
Fixes #ISSUE_Number
What does this PR do?
Type of Change
Breaking Changes
Test Plan
make installcheckmake -C src/test installcheck-cbdb-parallelImpact
Performance:
User-facing changes:
Dependencies:
Checklist
Additional Context
CI Skip Instructions