Difference in support for partitions between hive and Unity

[m208205]

The Unity migration guide (https://docs.databricks.com/en/data-governance/unity-catalog/migrate.html#before-you-begin) states the following:

Unity Catalog manages partitions differently than Hive. Hive commands that directly manipulate partitions are not supported on tables managed by Unity Catalog.

I'm unable to find documentation that elaborates on this statement.  What specifically is different in the handling of partitions between Unity and hive?  We have existing partitioned delta tables in our hive meta store that we are considering migrating to Unity.

[Kaniz_Fatma]

Hi @m208205, The Unity Catalog and Hive handle partitions differently, and this distinction is important when migrating existing partitioned Delta tables from Hive to Unity Catalog.

Here are the key differences:

1. Partition Management:

   • Unity Catalog: When using Unity Catalog, you won’t be able to directly manipulate partitions using Hive commands. Unity Catalog manages partitions in its own way.
   • Hive: In Hive, you can use commands like ALTER TABLE ... ADD PARTITION or ALTER TABLE ... DROP PARTITION to modify partitions directly.

2. Table History Migration:

   • When you run CREATE TABLE CLONE to migrate tables from Hive to Unity Catalog, the table history is not migrated. Any tables cloned to Unity Catalog are treated as new tables.

3. Managed Tables in Unity Catalog:

   • It’s recommended to create managed tables in Unity Catalog. Managed tables are the preferred way to create tables because they provide better control over metadata and lifecycle management.

For more detailed information, you can refer to the official documentation on upgrading Hive tables and views to Unity Catalog. This guide will help you navigate the migration process effectively. If you have any further questions, feel free to ask! 😊