pan3793 opened a new pull request, #18749: URL: https://github.com/apache/iceberg/pull/18749
Adds a Spark 4.3 integration (iceberg-spark-4.3_2.13, extensions, runtime) built against Spark 4.3.0 and makes 4.3 the default Spark version; 4.2 remains and is untouched. The series follows the usual approach for a new Spark version: the first two commits move spark/v4.2 to spark/v4.3 and copy it back, so the review is in the third. Its non-mechanical part adapts to Spark 4.3: the new nanosecond timestamp accessors on InternalRow and ArrayData, ParserInterface.splitStatements, and API moves in CoalesceExec and the V2 bucketing configs. The Spark module does not map nanosecond timestamps yet, so those accessors return micros with a zero sub-microsecond remainder. Test plan: 1. `./gradlew build -x test -x integrationTest -DsparkVersions=4.3` 2. `./gradlew :iceberg-spark:iceberg-spark-4.3_2.13:test :iceberg-spark:iceberg-spark-extensions-4.3_2.13:test` 3. `./gradlew :iceberg-spark:iceberg-spark-runtime-4.3_2.13:shadowJar` 4. Spark CI now includes 4.3 on JDK 17 and 21 (Scala 2.12 excluded); 4.2 jobs unchanged --- **AI Disclosure** - Model: DeepSeek V4.1 Flash - Platform/Tool: DeepSeek Harness - Human Oversight: fully reviewed - Prompt Summary: Port Iceberg's Spark integration from 4.2 to 4.3 following the move-then-copy-back approach, wire the new modules into the build, JMH and CI, and adapt code and tests to Spark 4.3 API changes. -- 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]
