Limitations and considerations

The following limitations apply when using Databricks Lakehouse (Delta):

• Supports LOB replication in Limited LOB mode only.
• Table names can only contain alphanumeric and underscore characters.

• Proxy limitations: See Setting advanced connection properties.

• General DDL limitations and considerations:
  • Dropping columns is not supported. Capturing changes to a table with a dropped column will set NULL values in the corresponding target column.
  • Renaming columns is not supported.
  • Modifying columns is not supported.
  • Renaming tables is not supported when the Storage type is Amazon S3.
• DDL limitations and considerations for the Apply Changes DDL Handling Policy:
  • When set to Ignore <DDL Type>:
    • When renaming or modifying a column, capturing changes will set NULL values in the corresponding target column.
    • Although truncating tables is supported, capturing changes from a truncated table will cause data conflicts as the target table will not be truncated.
    • Renaming tables is not supported. Capturing changes from a renamed table will result in the table being suspended (when the default table error handling policy is set). However, the table will be suspended only after the task is topped and resumed.
  • When set to <Action> target table:
    • When renaming or modifying a column, the table will be suspended.

• Target tables with buckets or partitions will not be created automatically on the target. Preexisting Databricks Lakehouse (Delta) target tables with buckets or partitions (which are identical to those of the corresponding source tables) are supported though.

• The Transactional apply Change Processing mode is not supported.

• The Apply Changes replication mode supports tables with a Primary Key/Unique Index only. Tables without Primary Keys/Unique Indexes will be suspended at the Change Processing stage.

  Information note

  Possible workarounds for tables without Primary Keys/Unique Indexes:

  • Configure a transformation to add a Primary Key/Unique Index column during replication
  • Create a Store Changes replication task for tables without Primary Keys/Unique Indexes
• The Store task recovery data in target database Change Processing option is not supported.
• The Replication Status Control Table (Name on target: attrep_status) is not supported.
• If a connectivity-token to Databricks expires during a task, you must stop and resume the task once the new token is set.
• Changing the Target table schema is the equivalent of changing the specified database name in the endpoint settings.