[arrow] Read inline VECTOR columns via Arrow FixedSizeList (#410 PR 2)#412
Draft
JunRuiLee wants to merge 4 commits into
Draft
[arrow] Read inline VECTOR columns via Arrow FixedSizeList (#410 PR 2)#412JunRuiLee wants to merge 4 commits into
JunRuiLee wants to merge 4 commits into
Conversation
JunRuiLee
force-pushed
the
feat/vector-type-pr2
branch
2 times, most recently
from
e889567 to
95566c3
Compare
Add VectorType (VECTOR<element, length>) as a standalone fixed-size dense vector type: u32 length validated to [1, i32::MAX], element type restricted to BOOLEAN/TINYINT/SMALLINT/INT/BIGINT/FLOAT/DOUBLE, JSON serde matching the Java wire shape, and a VECTOR<ELEM, N> Display.
Add the DataType::Vector variant and fill every exhaustive match: type-system semantics (is_nullable, copy_with_nullable, contains_row_type) implemented, and IO/integration sites returning explicit Unsupported (PR 1 supports the type in schemas; vector data IO lands in a later PR). Deserialize is hand-written to compose with the untagged DataType enum.
JunRuiLee
force-pushed
the
feat/vector-type-pr2
branch
from
95566c3 to
c1d8c2b
Compare
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
PR 2 of reading Paimon
VECTORcolumns in paimon-rust, tracked in #410. Enables reading vector columns stored inline in ordinary parquet data files.Roadmap (#410): PR 1 type system (#411) → PR 2 (this) Arrow conversion + parquet inline read → PR 3 dedicated
.vector.files.Changes
VectorType↔ ArrowFixedSizeListconversion inarrow/mod.rs(both directions), replacing PR 1'sUnsupported: forwardVector→FixedSizeList(Field("element", elem, element_nullable), length); reverse guards non-positive sizes withu32::try_fromand routes throughVectorType::try_new(rejecting length 0 and invalid element types).paimon_type_to_arrow, so an inlineFixedSizeListvector column now materializes directly.Unchanged boundaries (vectors still not filter-pushable / not cast / not partition or BTree keys / no
binary_rowdatum); writing and dedicated.vector.files remain out of scope.Testing
cargo test -p paimon --lib901 passed;cargo fmt --all -- --checkclean;cargo clippy -p paimon --all-targets --features fulltext,vortex,mosaic -- -D warningsclean. Adds 5 Arrow conversion unit tests + an end-to-end parquet test reading aFixedSizeList<Float32>column with a null row (downcast assertsvalue_length, child values, null bitmap).