What causes PROTOCOL_NOT_SUPPORTED?

The Delta table uses a newer feature (Deletion Vectors, Column Mapping, Liquid Clustering) that requires a higher protocol version than the client supports. An older Databricks Runtime or open-source Spark version is trying to read a table written by a newer runtime. A third-party tool (pandas, polars, DuckDB, Power BI Direct Lake) is reading a Delta table with protocol features it has not implemented. An upgrade to the Unity Catalog or Delta table itself bumped the protocol beyond what legacy clusters support

How do I fix PROTOCOL_NOT_SUPPORTED?

Upgrade the Databricks Runtime on the failing cluster to a version that supports the required protocol features.. For external tools, check whether the Delta Connector version for that tool supports the required protocol (e.g., delta-rs, delta-spark).. Check the required versions: `DESCRIBE DETAIL ` → `minReaderVersion`, `minWriterVersion`.. If Deletion Vectors are the only incompatible feature and they are not needed, they can be disabled: `ALTER TABLE t SET TBLPROPERTIES ('delta.enableDeletionVectors' = false)`.. For read-only external access, export the data to Parquet or use Delta Sharing instead of direct path access.

High severitysql

Power BI Refresh Error:
PROTOCOL_NOT_SUPPORTED

What does this error mean?

The client (Spark version, Databricks Runtime, or external reader) does not support the protocol version required by the Delta table. The table requires a minimum reader or writer version that exceeds what the client implements.

Common causes

1The Delta table uses a newer feature (Deletion Vectors, Column Mapping, Liquid Clustering) that requires a higher protocol version than the client supports
2An older Databricks Runtime or open-source Spark version is trying to read a table written by a newer runtime
3A third-party tool (pandas, polars, DuckDB, Power BI Direct Lake) is reading a Delta table with protocol features it has not implemented
4An upgrade to the Unity Catalog or Delta table itself bumped the protocol beyond what legacy clusters support

How to fix it

1Upgrade the Databricks Runtime on the failing cluster to a version that supports the required protocol features.
2For external tools, check whether the Delta Connector version for that tool supports the required protocol (e.g., delta-rs, delta-spark).
3Check the required versions: `DESCRIBE DETAIL <table>` → `minReaderVersion`, `minWriterVersion`.
4If Deletion Vectors are the only incompatible feature and they are not needed, they can be disabled: `ALTER TABLE t SET TBLPROPERTIES ('delta.enableDeletionVectors' = false)`.
5For read-only external access, export the data to Parquet or use Delta Sharing instead of direct path access.

Frequently asked questions

Can I downgrade a Delta table's protocol version?

Only if the feature requiring the higher version is disabled. For example, disabling Deletion Vectors allows dropping the minWriterVersion. Use ALTER TABLE SET TBLPROPERTIES to adjust the minimum versions after disabling features.