Shahzeb-DataEngineer opened a new pull request, #18836: URL: https://github.com/apache/iceberg/pull/18836
Documents the `NullabilityHolder.reset()` contract reported in #15808: `reset()` clears only the null count, leaving markers from the previous batch in place, so callers must set the marker for every index before reading it, as the vectorized readers do per batch. This follows the approach the reviewer suggested on #16264 and revives the stale-closed #17252, adding regression tests that lock the contract in. Happy to switch to clearing the markers in `reset()` instead if maintainers prefer that fix. Closes #15808. Test plan: new `TestNullabilityHolder` covers reset clearing the null count while preserving markers, and a full marker rewrite after reset describing the new batch. Verified with the project's google-java-format 1.22.0 (clean); the new tests pass (2/2), compiled and run directly with javac/JUnit because the Gradle daemon could not start in this environment. --- **AI Disclosure** - Model: Muse Spark 1.3 - Platform/Tool: Muse - Human Oversight: [unknown - human to fill in] - Prompt Summary: Document the `NullabilityHolder.reset()` null-marker contract and add regression tests. -- 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]
