sandip-db opened a new pull request, #51182: URL: https://github.com/apache/spark/pull/51182
### What changes were proposed in this pull request? The current Hadoop version in Spark hasn’t been compiled with Zstd because of which ZSTD compressed json, csv, txt and xml files cannot be read. Spark has a built-in `ZStdCompressionCodec`, which is used for compressing shuffled data and other use cases. This PR adds support for ZStandard decompression in file data source readers using the built-in `ZStdCompressionCodec`. It also add support for some non-standard extensions like `.gzip` and `.zstd`. The first commit in this PR is a vanilla copy of Hadoop's `LineRecordReader`. This was needed to add support for non-Hadoop based codec. Reviewers need not review this commit. Following config have been added to enable/disable changes in this PR: - `spark.sql.execution.datasources.hadoopLineRecordReader.enabled`: Setting this SQLCOnf to false will result in using Hadoop's `LineRecordReader`. Otherwise, use the inlined `HadoopLineRecordReader` by default. - `spark.io.zStandard.enabled`: Enable Spark's `ZStdCompressionCodec` for file data source reader. Default: `true` ### Why are the changes needed? Same as above. ### Does this PR introduce _any_ user-facing change? Yes, Adds support for ZSTD decompression in file data source readers. ### How was this patch tested? Added new unit tests ### Was this patch authored or co-authored using generative AI tooling? 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]
