Gabriel39 opened a new pull request, #66302: URL: https://github.com/apache/doris/pull/66302
## Changes - Backport the Iceberg Variant Parquet reader from #66206 to `branch-4.1`. - Read Iceberg Variant through File Scanner V2 while explicitly rejecting Variant in Avro, ORC, and the legacy scanner path. - Preserve shredded Variant columns in `ColumnVariantV2` and project typed leaves directly without canonical Variant materialization. - Push typed-leaf access paths into Parquet physical projection so page statistics can filter shredded leaf pages. - Add Iceberg Variant coverage for expressions, schema evolution, time travel, scanner compatibility, and unsupported write/format paths. - Adapt the callback and access-path APIs to `branch-4.1`, including preservation of subcolumn paths across derived Variant projection slots. ## Why Iceberg Variant files must retain their physical shredding benefits when read by Doris. Materializing every row into canonical Variant bytes would lose zero-copy typed-leaf access and prevent Parquet page-level pruning. ## Impact The change is limited to Iceberg Variant reads through the Parquet File Scanner V2 path. Iceberg Variant writes and Variant reads from Avro, ORC, or the legacy scanner remain unsupported and return an explicit error. ## Checks - BE targeted unit tests: 91 tests from 7 suites passed. - FE targeted unit tests: 60 tests passed. - Clang-format 16 dry run passed for all changed C/C++ files. - `git diff --check` passed. -- 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]
