Github user rdblue commented on the issue:
https://github.com/apache/spark/pull/23086
> I still do not think we should mix the catalog support with the data
source APIs
We are trying to keep these separate. `Table` is the only overlap between
the two. If you prefer more separation, we could move the `newScanBuilder`
method to a separate interface that readable data sources implement.
> Catalog is a well-defined concept in database systems, as what Spark SQL
follows. The so-called "table catalog" is not a catalog to me.
I'm glad that you're interested in joining the discussion on multi-catalog
support. Let's have that discussion on the catalog issues or discussion threads
on the dev list, not here on an update to the read API.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
