Github user mallman commented on the issue: https://github.com/apache/spark/pull/22905 > is there anything blocked by this? I agree this is a good feature, but it asks the data source to provide a new ability, which may become a problem when migrating file sources to data source v2. This isn't blocking anything. It's just a contribution that's shown itself to be very helpful for us identifying the source of performance problems in past experience. I think it would be helpful for others, too. That being said, I don't know enough about what would be involved in migrating file sources to data source v2 to say how difficult that would be. This implementation (for Parquet) is essentially a one-liner. All the heavy lifting is done by `SparkToParquetSchemaConverter`.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org