AnishMahto opened a new pull request, #58973: URL: https://github.com/apache/spark/pull/58973
This is a stacked PR. See incremental diff here: https://github.com/AnishMahto/spark/compare/scd1-leaf-versioning-dispatch...scd1-per-leaf-metadata-schema. See previous PR in stack here: https://github.com/apache/spark/pull/58915 ### Why are the changes needed? In order to support ignore-null in SCD1, we need to persist additional metadata per key that tells us the last event that authored each key, if any so far. This is effectively per-leaf level versioning, rather than the typical row/key-level versioning that SCD1 uses when ignore-null is off (the default case). The latter relies on the single upsert/delete version per row, also in the CDC metadata column. This is the SCD1 equivalent change of https://github.com/apache/spark/pull/58483. ### Does this PR introduce _any_ user-facing change? Yes. After this change existing AutoCDC tables will undergo nested schema evolution to gain a null version map field in their `cdc_metadata` column, when they are not using ignore-null (the default case). ### How was this patch tested? Added schema evolution test to `AutoCdcScd1SchemaEvolutionSuite`. Updated existing tests to reflect the additional version map field in `cdc_metadata`. ### Was this patch authored or co-authored using generative AI tooling? Co-authored with GPT 5.6 Sol. -- 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]
