mrhhsg opened a new pull request, #68125:
URL: https://github.com/apache/doris/pull/68125
### What problem does this PR solve?
Issue Number: None
Related PR: #67774
Problem Summary:
`__DORIS_VERSION_COL__`, `__DORIS_COMMIT_TSO_COL__`, and
`__DORIS_BINLOG_TSO__` can store placeholder values in segment files while
readers synthesize their logical values at read time. Segment-level expression
ZoneMap evaluation and page-level ZoneMap/Bloom filter pruning could evaluate
those physical placeholders and incorrectly discard rows that match the logical
values.
This change centralizes the lookup of read-time hidden-column constants.
Segment predicates and expression ZoneMap filters evaluate a synthetic
single-value ZoneMap, while page-level pruning skips the physical ZoneMap and
Bloom filter indexes because every page observes the same read-time value.
Row-level predicate evaluation remains unchanged.
### Release note
Fix incorrect pruning when filters reference read-time hidden version or TSO
columns.
### Check List (For Author)
- Test: Unit Test
- `BUILD_TYPE=ASAN ./build.sh --be -j 32`
- `BUILD_TYPE=ASAN ./run-be-ut.sh --run
--filter=SegmentIteratorExprZonemapTest.* -j 32`
- `./build-support/clang-format.sh`
- `./build-support/check-format.sh`
- `./build-support/run-clang-tidy.sh --base HEAD --build-dir
be/ut_build_ASAN`
- Behavior changed: Yes. Hidden-column predicates now use their logical
read-time values instead of physical placeholders for index pruning.
- Does this need documentation: No
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
