rdblue commented on a change in pull request #25348: [SPARK-28554][SQL] Adds a
v1 fallback writer implementation for v2 data source codepaths
URL: https://github.com/apache/spark/pull/25348#discussion_r315446002
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/DataSourceV2Strategy.scala
##########
@@ -200,24 +202,37 @@ object DataSourceV2Strategy extends Strategy with
PredicateHelper {
catalog,
ident,
parts,
+ query,
planLater(query),
props,
writeOptions,
orCreate = orCreate) :: Nil
}
case AppendData(r: DataSourceV2Relation, query, _) =>
- AppendDataExec(r.table.asWritable, r.options, planLater(query)) :: Nil
Review comment:
I like the idea to use the same physical plan node when the semantics match,
but I'm not sure that it is possible to make strong guarantees about not
changing v1 if we do. The problem is that there are multiple v1 plan nodes for
the same operation, which could have slightly different behavior.
We could take the time to inspect the v1 implementations and convert, but
that adds risk and takes time. It also isn't needed to migrate to v2, and
wouldn't speed up the migration, unless I'm missing something. So it probably
doesn't provide enough value to make doing it worth while.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
