Read manifests with V3 projection - #3690
Merged
Merged
Conversation
Read manifest entries and manifest lists with the latest supported schema so V3-only fields are retained while older manifests resolve missing fields to null.
Exclude V3-only properties from V1 and V2 fastavro comparisons and clarify the latest read-version constant name.
Exclude V3-only properties from the reflected V2 manifest dictionary used by the REST integration test.
Fokko
reviewed
Aug 2, 2026
Fokko
left a comment
Fokko
left a comment
Contributor
There was a problem hiding this comment.
I think the right approach would be to bump DEFAULT_READ_VERSION instead of introducing _LATEST_MANIFEST_READ_VERSION. Having the two introduces complexity that we want to avoid. When we have the versions aligned, then the projection should populate all the fields (and fill in None for ≤V2).
kevinjqliu
self-requested a review
August 4, 2026 16:25
Contributor
Author
|
Hi @Fokko @kevinjqliu would you mind taking another look when you have time? Thanks! |
winding-lines
added a commit
to magmalake/magmalake.org
that referenced
this pull request
Aug 31, 2026
The correctness section claimed this work "surfaced upstream bugs in PyIceberg, pyarrow and iceberg-rust" without saying which, or where. Each disagreement has now been reduced to a standalone reproducer with no magmalake code in it. Two turned out to be unreported and were filed: - apache/arrow#51097 — the Parquet writer undercounts nulls for a fixed-width leaf under list<struct> when some lists are null or empty, while the BYTE_ARRAY leaf beside it is correct. - apache/iceberg-rust#3118 — an In predicate whose literals straddle a file's bounds is never pruned, because each bound is tested against the whole literal set rather than narrowing it. Three were already open upstream and are linked rather than duplicated: apache/iceberg-python#3690, #3620 and #3833. Two earlier candidates are dropped from the list. pyarrow reporting LZ4_RAW as "LZ4" and deriving no ConvertedType for a non-UTC timestamp are both deliberate API behaviour, not defects. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_015or6oTZmb7EQTXTbALbBXY
Fokko
approved these changes
Aug 31, 2026
Fokko
left a comment
Fokko
left a comment
Contributor
There was a problem hiding this comment.
Clean and crisp, I love it! Thanks @KaiqiJinWow 🙌
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
2 participants
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.
Rationale for this change
V3 manifest schemas are defined, but manifest entries and manifest lists are still read with the V2 projection. Avro schema resolution therefore drops V3-only fields such as
first_row_id,referenced_data_file,content_offset, andcontent_size_in_bytesbefore they reachDataFileandManifestFile.Use V3 as the default read and in-memory projection. V1 and V2 writers continue to use their version-specific file schemas, and older manifests remain compatible because missing V3 fields resolve to null.
This isolates the read-side concern raised in #3624 from its V3 writer work and is a prerequisite for #3478 to consume deletion-vector content ranges from catalog manifests.
Are these changes tested?
Yes.
first_row_id.pytest tests/avro/test_file.py tests/utils/test_manifest.py: 50 passed.datafusionandpyiceberg-coreextras were excluded.Are there any user-facing changes?
No. This fixes internal manifest deserialization and prepares V3 read support without changing behavior for currently supported V1 and V2 tables.