nssalian opened a new pull request, #19032: URL: https://github.com/apache/iceberg/pull/19032
Backports https://github.com/apache/iceberg/pull/19009 to Spark 4.0 and 4.1. Part of https://github.com/apache/iceberg/issues/18669. `SparkBatch` no longer sends variant scans to the row reader because of the shred table property, `none` or `counts` metrics, or variant lower bounds, and `SparkScanBuilder` no longer loads variant stats for that check. https://github.com/apache/iceberg/pull/18851 already reads shredded files per file in the vectorized reader. Disabling vectorization still selects the row reader. Test plan: - `TestSparkVariantRead` in 4.0 and 4.1 is identical to the 4.2 version from https://github.com/apache/iceberg/pull/19009. The same tests fail without this change in both versions. --- **AI Disclosure** - Model: Claude Opus 5.5 - Platform/Tool: Claude Code - Human Oversight: fully reviewed with tests run locally - Prompt Summary: Backport the removal of the Spark plan-time shredded variant checks to Spark 4.0 and 4.1, with the same Spark tests. The code changes and tests are generated. -- 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]
