costas-db opened a new pull request, #3827:
URL: https://github.com/apache/parquet-java/pull/3827
### Rationale for this change
Parquet column paths are component-based, but Bloom filters are currently
stored under dot-string keys. A top-level field named `a.b` (`["a.b"]`)
therefore collides with nested field `a`.`b` (`["a", "b"]`).
This draft is reproduction-only for #3826. It intentionally does not include
a production fix.
### What changes are included in this PR?
- Adds a direct regression that writes distinct Bloom-filter values to the
colliding paths and checks each footer column for its own values.
- Adds a small expected-output JSON fixture that makes the failure readable
without replacing the direct assertions.
### Are these changes tested?
The focused test reaches the intended assertion and fails on current
`master`:
```text
expected:
"topLevelContainsOwnBloomValues" : true
actual:
"topLevelContainsOwnBloomValues" : false
```
The nested path remains `true`, demonstrating that its Bloom filter
overwrote the top-level filter under the shared `"a.b"` key.
`spotless:check` and `apache-rat:check` pass.
### Are there any user-facing changes?
No. This draft only adds a reproduction.
Relates to #3826.
--
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]
