XiaoHongbo-Hope opened a new pull request, #1034:
URL: https://github.com/apache/paimon-rust/pull/1034

   ### Purpose
   
   Allow native Parquet writes to tune dictionary encoding for high-cardinality 
columns without disabling it for other columns. Java already exposes these 
configuration keys; the Rust writer previously kept parquet-rs defaults 
regardless of these options.
   
   ### Changes
   
   - Honor `parquet.enable.dictionary` and 
`parquet.enable.dictionary#column.path`.
   - Preserve the default and give physical leaf overrides precedence over the 
global setting.
   - Validate boolean values and empty path components before opening the 
output file.
   - Use the common Parquet writer for ordinary and shredded data. Paths refer 
to physical leaves, not logical MAP keys; existing files are unchanged.
   
   ### Validation
   
   Tests inspect actual Parquet dictionary pages and round-trip values for 
global/per-column overrides and nested fields. Invalid configurations must not 
create files. Performance gains depend on data and require rewriting; this PR 
does not change defaults or claim a universal throughput gain.
   
   - Parquet tests: 84 passed.
   - MAP-related tests: 117 passed, 1 ignored (overlaps the Parquet suite).
   - `cargo fmt --all -- --check` and `git diff --check`: passed.
   


-- 
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