Repository navigation
Conversation
… deletes - LLAP reports the file row of the first row of each batch: ORC files with position deletes or virtual columns read through LLAP IO - The non-LLAP vectorized ORC reader forces orc.sarg.to.filter off: the row-level filter drops rows and whole batches without reporting their positions; the SARG still prunes stripes and row groups (HIVE-30110) - Impact: correctness/routing; ORC files with deletes now read through LLAP IO (not timed)
asf-ci-hive
added
tests unstable
tests pending
and removed
tests pending
tests unstable
labels
- llap_iceberg_read_orc_deletes.q: V2 position deletes and V3 deletion vectors on an ORC file with three row groups, read through LLAP IO (LLAP IO counters in the golden) - The row-mode ORC reader forces orc.sarg.to.filter off too: Iceberg's OrcIterable takes the batch's file position before ORC can skip filtered batches - LlapRecordReader.getRowNumber returns Long.MIN_VALUE (unknown) instead of throwing - OrcEncodedDataConsumer: drop the redundant noIndex check - TestOrcEncodedDataConsumerRowPositions: drop redundant tests and asserts; a read without columns replaces the faked ALL_RGS case - HiveVectorizedReader comments
asf-ci-hive
added
tests unstable
tests pending
and removed
tests pending
tests unstable
labels
deniskuzZ
marked this pull request as ready for review
October 29, 2026 10:37
- ColumnVectorBatch: startRowInFile is private, with accessors - LlapRecordReader is final - OrcEncodedDataConsumer: row group start row without a nested ternary - TestOrcEncodedDataConsumerRowPositions: Hive's ASF header; SerDeStripeMetadata and Mockito mocks instead of hand-written fakes; one row batch per stripe in the fixture - IcebergRecordReader: sarg-to-filter set in the ORC read builder chain, next to the filter - IDE inspection cleanups in the touched classes
asf-ci-hive
added
tests passed
tests pending
and removed
tests pending
tests passed
labels
Contributor
Member
Author
|
those are pre-existing sonar issues in origin/master, but i'll handle them |
- LlapRecordReader: merge the nested decimal64 if - OrcEncodedDataConsumer: braces on the null-reader continue
|
asf-ci-hive
added
tests unstable
tests pending
and removed
tests pending
tests unstable
labels
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
ColumnVectorBatch.startRowInFile), from the stripe start rows and the row group index; LlapRecordReader exposes it asRowPositionAwareVectorizedRecordReader.getRowNumbergetRowNumberreturnsLong.MIN_VALUE(unknown), and the delete filter fails if it needs a position instead of applying deletes to the wrong rowsorc.sarg.to.filteroff, vectorized (non-LLAP) and row mode: ORC's row-level filter drops rows and whole batches without reporting their positions, which position deletes, ROW__POSITION and row lineage take from the batch index; the SARG still prunes stripes and row groups (HIVE-30110)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?
orc.sarg.to.filteris ignored for Iceberg ORC scans, vectorized and row mode; the SARG still prunes stripes and row groups. No other user-facing changeHow was this patch tested?