Extend transformed query reduction to DML - #1362
Closed
tlmorgan24 wants to merge 8 commits into
Closed
tlmorgan24 wants to merge 8 commits into
tlmorgan24 wants to merge 8 commits into
Conversation
… (e.g. incorporate rowId setup) to make it easier to verify the logs independently
…se-reducer was enabled
…d as diff rows in logs
generateUnoptimizedQueryString projected COUNT(*) and cleared the WHERE clause without re-projecting the predicate, so the unoptimized side counted every row of the table regardless of the condition. Every run whose predicate filtered anything therefore reported a mismatch. Project the predicate per row instead, as the other NoREC generators do. A CASE expression is used rather than a cast because HSQLDB does not allow casting a BOOLEAN to an integer. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01G1XqNaswnH8mWTXFkkGa1X
generateOptimizedQueryString projected a literal COUNT(*) column in its non-aggregate branch as well as in its aggregate one. NoRECOracle counts the rows that branch returns, and an aggregate always returns exactly one, so the optimized count was 1 whatever the predicate matched. Project * in the non-aggregate branch, as PostgresExpressionGenerator does. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01G1XqNaswnH8mWTXFkkGa1X
generateOptimizedQueryString set the WHERE clause only in its non-aggregate branch, so the aggregate branch counted every row of the table while the unoptimized side applied the predicate. Set the WHERE clause for both branches. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01G1XqNaswnH8mWTXFkkGa1X
The expression generator can produce chr() with a negative argument, on which YSQL raises "character number must be positive". That error was missing from YSQLErrors, so the oracle reported it as an unexpected error rather than ignoring the statement. PostgresCommon already lists it (PostgresCommon.java:87); YSQLErrors otherwise mirrors that list. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01G1XqNaswnH8mWTXFkkGa1X
Collaborator
Author
|
By solving the bug where an oracle would not report its findings, some pre-existing defects unrelated to my work were revealed in the CI tests. After fixing these too, the current PR is trying to accomplish too many things at once, so I am closing it and splitting it into two separate PRs. |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
The previously merged transformed query reduction (PR #1360) was for the EET SELECT oracle only. This extends the work to also benefit the EET DML oracle; the theoretical approach is the same (implementing the unchanged TransformationReproducer interface). The main difference is that a DML statement's transformable expressions vary by statement kind (e.g. UPDATE has SET values while DELETE does not), so much of the diff is to make the reproducer hold them in a fixed order with a reassembly function that rebuilds the statement string.
Other parts of the diff ensure the EET DML logs clearly lay out the reduction. A pre-existing bug is also solved, where any oracle supplying a reproducer would only report its findings when --use-reducer was enabled.