Improve how pruning estimates how much of a table will be removed #6012
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
e4b8b3c to
bb2c164
Compare
bb2c164 to
6ee64b4
Compare
zorancv
reviewed
May 16, 2025
zorancv
approved these changes
May 19, 2025
Our current estimate is way too naive and causes a lot of unnecessary table rebuilds. Using statistics from pg_stats should improve that
6ee64b4 to
6cd68d8
Compare
We are not using those right now and there's therefore no point in storing them
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Pruning can use one of two strategies to get rid of unneeded rows in a table: deleting or rebuilding. Rebuilding is better when we remove more than 50% of a table, otherwise, deletion should be used.
Currently, we estimate how much of a table will be removed by looking at the percentage of historical blocks we prune weighted by the ratio of entity versions to entities. This basically assumes that entity changes are very evenly distributed across blocks. In practice, this leads to serious overestimation of how much of a table will be removed. In one case, the estimate was that 58% of a table would be removed, when in reality only 2% of the table would be removed.
With this change, we use database statistics to improve the estimates which reflect reality much more closely.