Not able to create external table In hive_metastore permanent
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-07-2025 08:51 PM
Hi @everyone i am using databricks community edition in which i created one database and some external table in the hive_metastore catalog but i terminate the spark cluster then i came back so there was no table could anybody help me with this why its not showing me ?
i created this way:-
here you can see there is no landing_club database ? why it is happening ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-08-2025 05:42 AM
I think hive_metastore only stores metadata. And Hive registers metadata about the table and its location, but the data files are not managed by Hive.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-14-2025 06:10 AM
And the community edition, all the created tables would be deleted once the compute being terminated. Because the hive metastore is local to compute. Means like a memory as RAM. Once erased, everything would be gone. As community designed as learn and experimentation usage (limited resource).
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-14-2025 09:54 AM
The tables created via Hive Metastore, stores data in dbfs://user/warehouse/.... but interestingly the Hive metastore by default is local, so it stores metadata related information in the DBFS (which is the attached cluster in Databricks) and since the table is external so data is persisted in cloud storage (S3). Hence deleting the cluster, terminates the Meta data like the Hive table but however since the table is External, so the data is persisted in S3 via DBFS.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
04-29-2025 07:31 PM
To view the tables that have been created, you may need to start the cluster first.