jbonofre opened a new pull request, #1474:
URL: https://github.com/apache/arrow-java/pull/1474

   ## What's Changed
   
   Follow-up to #1389, which this PR is based on: until #1389 is merged, only 
the last commit is new.
   
   - `JsonFileWriter` and `JsonFileReader` look each vector up by position 
instead of through `getVector(Field)`, so columns with equal Fields (such as 
duplicate names with the same type) no longer share one vector. A JSON round 
trip of `x: [1], x: [2]` read back `x: [null], x: [2]` before.
   - The Field lookup of `VectorSchemaRoot` keeps the first of the columns with 
equal Fields, as `getVector(String)` documents, instead of the last.
   - `syncSchema()` keeps the custom metadata of the schema.
   - Review comments on #1389: the constructor and `syncSchema()` share the 
code that builds the lookup, the comments say the old keys no longer match 
because the promoted Field is structurally different, an assertion that could 
not fail is dropped, and the JSON round trip after a schema sync moved to 
`TestJSONFile`.
   
   `TestJSONFile` covers a JSON round trip of duplicate columns through both 
`read()` and `read(VectorSchemaRoot)`, and `TestVectorSchemaRoot` covers the 
lookup of duplicate columns and the metadata after `syncSchema()`.
   
   Closes #1473.
   


-- 
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]

Reply via email to