thswlsqls opened a new issue, #10380: URL: https://github.com/apache/paimon/issues/10380
**Search before asking** - [x] I searched in the [issues](https://github.com/apache/paimon/issues) and found nothing similar. **Paimon version** master @ b0af45681 **Compute Engine** Engine-agnostic (core): Java `withFilter`, Spark full-text/hybrid search with `WHERE`. **Minimal reproduce step** 1. Data-evolution table, `scalar-index.search-mode=full` (or `detail`), full-text index on `content`, FM index on the filter column (`content`). 2. Full-text search with `withFilter(equal(content, 'paimon zeta'))`. 3. The result is empty. A plain table scan with the same filter returns the row. FM always declines `=`, so `DataEvolutionFullTextRead#matchedRows` gets no evaluation and drops the covered rows. A full-text index on the filter column behaves the same. **What doesn't meet your expectations?** In `full`/`detail`, these rows should be decided from their filter columns (`fast` unchanged): - full-text.mdx: `full`/`detail` read filter columns of unindexed rows. - fm.mdx: a declined predicate falls back to the data path. - `DataEvolutionGlobalIndexScanner` already treats runtime declines as unindexed ranges (#10339). - The `fast` warning recommends `full`, which currently returns nothing. **Anything else?** N/A **Are you willing to submit a PR?** - [x] I'm willing to submit a PR! -- 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]
