deniskuzZ opened a new pull request, #6920: URL: https://github.com/apache/hive/pull/6920
### What changes were proposed in this pull request? - HiveVectorizedReader no longer bypasses LLAP IO for an ORC file with position deletes or projected virtual columns - OrcEncodedDataConsumer reports the file row of each batch's first row (`ColumnVectorBatch.startRowInFile`), from the stripe start rows and the row group index; LlapRecordReader exposes it as `RowPositionAwareVectorizedRecordReader.getRowNumber` - GenericColumnVectorProducer (re-encoded input) has no file positions; asking it for one fails instead of returning a wrong position - Based on HIVE-30110 (#6839), which this PR includes until it is merged: positions assume batch row i is file row start + i ### Why are the changes needed? Such files were read around LLAP IO, without its cache, because LLAP could not report row positions ### Does this PR introduce _any_ user-facing change? No ### How was this patch tested? TestOrcEncodedDataConsumerRowPositions: row group positions across stripes, batches within a row group, files without a row index, ALL_RGS, no stripes -- 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]
