Hi Pralabh, If it helps, it is probably related to this change https://github.com/apache/spark/pull/28527
Regards On Mon, Jan 10, 2022 at 10:42 AM Pralabh Kumar <pralabhku...@gmail.com> wrote: > Hi Spark Team > > When creating a database via Spark 3.0 on Hive > > 1) spark.sql("create database test location '/user/hive'"). It creates > the database location on hdfs . As expected > > 2) When running the same command on 3.1 the database is created on the > local file system by default. I have to prefix with hdfs to create db on > hdfs. > > Why is there a difference in the behavior, Can you please point me to the > jira which causes this change. > > Note : spark.sql.warehouse.dir and hive.metastore.warehouse.dir both are > having default values(not explicitly set) > > Regards > Pralabh Kumar >