Github user rdblue commented on the issue:
https://github.com/apache/spark/pull/21306
@cloud-fan, @gatorsmile, I don't think this should be merged yet.
I've been implementing CTAS and RTAS based on this commit and I don't think
it makes sense to get a `TableCatalog` instance from the data source. The data
source should be determined by the catalog, not the other way around.
Otherwise, we could have a case where a `test` catalog uses the `parquet`
source, but that `parquet` source would return a `prod` catalog for its
`TableCatalog` because `prod` is the default. If catalogs can reuse data
sources, then the catalog should be determined first.
FYI @bersprockets, @felixcheung, @henryr
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
