AnuragRaut08 opened a new pull request, #25334:
URL: https://github.com/apache/datafusion/pull/25334

   ## Which issue does this PR close?
   
   * Closes #24776.
   
   ## Rationale for this change
   
   `ColumnRelation` currently round-trips `TableReference` qualifiers through a 
single dotted string. When a catalog, schema, or table identifier itself 
contains `.`, the qualifier boundaries are lost during deserialization, causing 
the reconstructed `TableReference` to differ from the original.
   
   For example, a schema named `my.schema` can be incorrectly interpreted as 
separate qualifier components after a protobuf round-trip.
   
   ## What changes are included in this PR?
   
   * Add a structured `parts` field to `ColumnRelation`.
   * Serialize `TableReference` components using `TableReference::to_vec()`.
   * Prefer the structured `parts` representation when deserializing.
   * Fall back to the existing `relation` string for legacy protobuf messages.
   * Add round-trip tests for dotted bare, partial, and full table references.
   * Add a test covering decoding of legacy relation-only messages.
   * Regenerate the protobuf bindings.
   
   ## What is the testing strategy for this PR?
   
   Added regression tests covering the `Column -> protobuf -> TableReference` 
round-trip for:
   
   * Bare table references containing `.`
   * Partial references with dotted schema identifiers
   * Full references with dotted schema identifiers
   * Legacy `relation`-only protobuf messages
   
   Also ran:
   
   `cargo test -p datafusion-proto-common column_relation -- --nocapture`
   
   `cargo test -p datafusion-proto-common`
   
   All tests pass.
   
   ## Are there any user-facing changes?
   
   No.
   


-- 
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]

Reply via email to