thswlsqls opened a new pull request, #10381: URL: https://github.com/apache/paimon/pull/10381
### Purpose fix #10380 - When every scalar index covering the filter columns declines the predicate (e.g. FM on `=`), `DataEvolutionFullTextRead#matchedRows` dropped the covered rows, so `full`/`detail` returned nothing. - In `full`/`detail` these rows are now decided by reading their filter columns, as full-text.mdx documents for unindexed rows and fm.mdx promises for declined predicates. - Mirrors the runtime-decline fallback of `DataEvolutionGlobalIndexScanner` (#10339). - Default `fast` is unchanged; extra reads happen only in opt-in `full`/`detail` and only for declined rows, the same cost as an unindexed column. ### Tests - Added `FullTextSearchBuilderTest#testFullTextSearchReadsFilterColumnsWhenScalarIndexDeclinesFilter` (`fast`/`full`/`detail`). Without the fix, `full`/`detail` fail (expected `[5]`, got `[]`); with it, 3/3 pass. - `mvn -pl paimon-core clean install` (JDK 11): all tests pass, including `FullTextSearchBuilderTest` 57/57; `PostgresqlCatalogTest` was skipped (needs Docker). -- 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]
