goutamadwant opened a new pull request, #12397: URL: https://github.com/apache/seatunnel/pull/12397
### Purpose of this pull request Related to #9746 and #9744. This implements the Spark slice; Flink remains separate. Fix nested-array conversion failures, preserve null map elements instead of converting them to empty maps, and avoid modifying caller-owned typed arrays. Separately, extend the schema conversion utility to handle recursively supported array element types using the existing SeaTunnel array API. ### Does this PR introduce _any_ user-facing change? Yes. Nested values now preserve correctly typed empty arrays and null elements across Spark source, transform and sink conversion boundaries. Primitive-array mappings remain unchanged. No configuration, dependency or public API signature changes are introduced. This does not add new SeaTunnel Sql syntax, change schema configuration syntax, or expand connector-specific type support. ### How was this patch tested? - Reproduced failures against unchanged converters. - Passed 63 regression and adjacent tests on both Java 8 and Java 11. - Covered nested arrays, maps, rows, decimals, binary values, dates, nulls and empty arrays. - Tested Catalyst unsafe projection and real Spark repartition/collect execution. - Passed the SeaTunnel FakeSource → Sql → Assert pipeline on Spark 2.4.6 and 3.3.0, checking exact nested values. - Spotless passed. ### Check list - [x] English and Chinese documentation updated. - [x] Regression and integration coverage added. - [x] No new dependencies or binary packages. - [x] No incompatible configuration or public API changes. - [x] New connector registration changes are not applicable. -- 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]
