Skip to content

feat(types): add Variant type with non-shredded codec and encryption-keys metadata#23

Open
abnobdoss wants to merge 4 commits into
mainfrom
v3/t6-variant-encryption
Open

feat(types): add Variant type with non-shredded codec and encryption-keys metadata#23
abnobdoss wants to merge 4 commits into
mainfrom
v3/t6-variant-encryption

Conversation

@abnobdoss

Copy link
Copy Markdown
Owner

No description provided.

Abanoub Doss added 4 commits June 23, 2026 08:40
Adds VariantType primitive (parses/serializes 'variant', min format v3) and
a pure-python encoder/decoder for the Parquet Variant binary format
(metadata dictionary + value: null/bool/int/double/string/object/array),
independent of PyArrow's still-unimplemented variant extension type
(apache/arrow#45937, #50131, #50132 all open).
…#2118)

Adds metadata-only encryption fields (no crypto/KMS): EncryptedKey model
(key-id, encrypted-key-metadata, encrypted-by-id, properties) matching the
Java EncryptedKeyParser; optional encryption-keys list on TableMetadataV3
and optional key-id on Snapshot. Read/deserialize + alias round-trip only;
V3 full-write still raises NotImplementedError upstream (apache#1551).
Place offset_size_minus_one in metadata header bits 6-7 and check the
reserved bit at bit 5, matching the Parquet Variant spec (previously bits
5-6 / bit 7, which only round-tripped with itself and rejected valid
metadata using a 3- or 4-byte offset size). Add decimal, date, timestamp,
binary and float primitive support, byte-level spec assertions, and
edge-case tests (empty object/array, large dictionary, float decode).
@abnobdoss abnobdoss closed this Jun 25, 2026
@abnobdoss abnobdoss reopened this Jun 25, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant