Gabriel39 opened a new pull request, #68582: URL: https://github.com/apache/doris/pull/68582
Ports #68568 to master, adapted to the Iceberg connector architecture. Master already uses `SchemaAwareDataTableScan` for historical schemas, but still rebinds every table partition spec. Renaming a column and then adding an unused partition field with that old name can make a historical query fail with `Cannot create identity partition sourced from different field in schema` even though the selected snapshot never uses the new spec. Bind only specs referenced by the selected snapshot's data and delete manifests to the selected scan schema. Use the same helper for SDK file planning, streaming file estimates, and manifest-cache planning. Retain the existing fast path when the scan uses the current schema. Adapt the historical scan tests to exercise partitioned/unpartitioned tables, rename/drop, schema-only/append, reused names, synchronous/streaming planning, and cache on/off. Assert that a nonmatching manifest in the selected snapshot is pruned and that cache planning succeeds without fallback. Also port the SQL snapshot/tag regression suite from #68568 unchanged. Validation: - The new metadata-only partition-evolution test reproduces the identity-partition error on unmodified master. - All 181 `IcebergScanPlanProviderTest` tests passed with the Maven build cache disabled. - Connector reactor compilation, Checkstyle, and the connector import gate passed. - The SQL regression suite is included; end-to-end SQL execution was not repeated on master in this port. -- 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]
