cloud-fan commented on pull request #28026: URL: https://github.com/apache/spark/pull/28026#issuecomment-625749455
> Are you saying that a Hive catalog plugin and table implementation would be forced to implement Spark's incompatible behavior somehow? I'm saying the Hive behavior about EXTERNAL doesn't make sense and Spark shouldn't allow catalog implementations to follow the Hive behavior. Besides, Spark provides an API and we must make sure the semantic is clear. It's bad to say: "I don't know what EXTERNAL means, ask the developer of your catalog." Spark has clearly defined the behavior of EXTERNAL for years and I don't agree we should allow catalogs to interpret it differently. Like the `USING provider`, it should only affect performance but not behaviors, whatever the catalog you use. ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
