What causes DELTA_LOG_CHECKPOINT_CORRUPTED?

A checkpoint write was interrupted mid-stream by a cluster failure, leaving a partial Parquet file. Cloud storage experienced a transient write error that silently corrupted the checkpoint. A direct file system operation deleted or truncated the checkpoint without updating the log. Storage-level bit rot or checksum errors corrupted the Parquet checkpoint file

How do I fix DELTA_LOG_CHECKPOINT_CORRUPTED?

Step 1: Identify the corrupted checkpoint by running DESCRIBE HISTORY table_name to find the last successful version.. Step 2: Delete the corrupted checkpoint file(s) from the _delta_log directory — Delta Lake will rebuild the state from individual JSON commit files.. Step 3: Run VACUUM RETAIN 0 HOURS on a copy of the table in a development environment to validate the log can be replayed.. Step 4: After deleting the corrupted checkpoint, run a read query to confirm the table is accessible.. Step 5: Enable Delta Lake redundancy or use cloud storage versioning to recover from future storage-level corruptions.

Critical severityexecution

Power BI Refresh Error:
DELTA_LOG_CHECKPOINT_CORRUPTED

What does this error mean?

Delta Lake found a corrupted or unreadable checkpoint file in the transaction log, preventing it from efficiently loading the table state.

Common causes

1A checkpoint write was interrupted mid-stream by a cluster failure, leaving a partial Parquet file
2Cloud storage experienced a transient write error that silently corrupted the checkpoint
3A direct file system operation deleted or truncated the checkpoint without updating the log
4Storage-level bit rot or checksum errors corrupted the Parquet checkpoint file

How to fix it

1Step 1: Identify the corrupted checkpoint by running DESCRIBE HISTORY table_name to find the last successful version.
2Step 2: Delete the corrupted checkpoint file(s) from the _delta_log directory — Delta Lake will rebuild the state from individual JSON commit files.
3Step 3: Run VACUUM RETAIN 0 HOURS on a copy of the table in a development environment to validate the log can be replayed.
4Step 4: After deleting the corrupted checkpoint, run a read query to confirm the table is accessible.
5Step 5: Enable Delta Lake redundancy or use cloud storage versioning to recover from future storage-level corruptions.

Frequently asked questions

How often does Delta Lake create checkpoints?

By default, Delta Lake creates a checkpoint every 10 commits. This is configurable via the delta.checkpointInterval table property.

Is it safe to delete a checkpoint file from _delta_log?

Yes — Delta Lake can reconstruct table state from the individual JSON commit files. Deleting a corrupted checkpoint forces Delta to fall back to replaying commits, which is slower but safe.