ad1happy2go opened a new pull request, #20523: URL: https://github.com/apache/hudi/pull/20523
### Describe the issue this Pull Request addresses closes #20522 Snapshot reads in the Trino connector fail on Merge-On-Read tables whose log files hold parquet data blocks (`hoodie.logfile.data.block.format=parquet`, table version 9 and below), which any Spark record merger writes. The file-group reader hands each block to the connector as an `inlinefs://` path, and `HudiTrinoReaderContext` rejected it with "Inline log blocks are not supported". ### Summary and Changelog A parquet data block is a complete parquet file embedded at a byte range of the log file, so the connector now reads it in place with the parquet page source it already uses for native log files. - `InlineTrinoInputFile`: a `TrinoInputFile` over one byte range of another file, presented as a file starting at position 0. Reads and footer (tail) reads stay inside the range. - `HudiPageSourceProvider`: log reads with an `inlinefs://` path open through that view of the log file. - `HudiTrinoReaderContext`: drops the rejection. - Tests: `TestInlineTrinoInputFile` covers the range mapping and bounds. `TestHudiParquetLogBlocks` reads a version 9 MOR table written with parquet log blocks, one record per block so a log file holds several, covering winning and losing updates, a soft delete, a hard delete, narrow projections and a filter. Its fixture fails if the log files do not hold parquet blocks. Without the fix the snapshot tests fail with the original exception. ### Impact Trino snapshot queries work on MOR tables with parquet log blocks. No config or API change. ### Risk Level low. The change only affects reads that previously failed; other log and base file reads take the same path as before. ### Documentation Update none ### Contributor's checklist - [x] Read through [contributor's guide](https://hudi.apache.org/contribute/how-to-contribute) - [x] Enough context is provided in the sections above - [x] Adequate tests were added if applicable -- 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]
