Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/22466 This is a behavior change and makes us different from Hive. However I can't find a strong reason to do it. It's like importing a database, but we can't automatically create table entries in the metastore when creating a database with an existing location. To me a more reasonable behavior is, fail earlier when creating a database with an existing and non-empty location.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org