Gabriel39 opened a new pull request, #68568: URL: https://github.com/apache/doris/pull/68568
Historical Iceberg predicates can still fail after #67479: Doris batch-mode detection and manifest-cache planning use table specs bound to the current schema, and Iceberg 1.11 skips rebinding when a schema-only update leaves the data snapshot unchanged. Bind partition specs by field ID to the schema selected by the Doris relation for both SDK and Doris planning. Preserve the existing scan context and reuse specs already bound to that schema. Tests now exercise Doris scan creation and batch-mode detection across rename/drop, schema-only/append, partitioned/unpartitioned tables, and reused field names. A new SQL regression covers snapshot and tag reads with batch mode and manifest caching enabled and disabled. Validation: - The four targeted tests fail on the unmodified branch-4.1 baseline with `Cannot find field`. - FE compilation and 147 Iceberg scan/utility unit tests passed. - The new regression suite passed against a local FE/BE and Iceberg REST/MinIO environment: 288 historical-query result assertions, no skipped tests. - FE Checkstyle passed with zero violations. -- 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]
