hello-stephen opened a new pull request, #65957:
URL: https://github.com/apache/doris/pull/65957
### What problem does this PR solve?
Issue Number: None
Related PR: #65660
Problem Summary:
PR #65660 was merged before its separate review follow-up could land. Two
correctness items therefore remain on master:
- doc-value offsets are cumulative, so checking a non-zero end offset can
let a previous row's doc snapshot hide sparse fields in the current row;
- two empty-object rows in the regression oracle were rendered as arrays,
which would lock unrelated corruption into the expected output.
This PR reapplies only the four-file review follow-up on top of current
master. It uses the current-row doc-value offset delta, adds the
prior-doc/current-sparse unit scenario, and removes the two unrelated oracle
rows.
### Release note
None
### Check List (For Author)
- Test:
- `build-support/clang-format.sh` and `build-support/check-format.sh` with
clang-format 16 passed.
- `git diff --check` passed.
- The changed Groovy case compiled with Groovy 4.0.19 and was discovered
as `test_topn_lazy_materialize_sparse_variant` in group `p0`.
- Full local regression framework packaging was attempted but stopped
before source compilation because the worktree dependency bundle does not
contain `thirdparty/installed/bin/thrift`; BE UT and buildall are required in
CI.
- Behavior changed: Yes. Current-row sparse fields are no longer hidden by a
previous row's doc snapshot.
- 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]
