Finding and exporting your Databricks data

What it holds

Delta tables hold structured records backed by data files and transaction metadata. Older versions may remain in storage until retention and cleanup policies remove them.

Where Databricks shows the size

1. Enumerate tables with SHOW TABLES IN my_catalog.my_schema;
2. Run DESCRIBE DETAIL my_catalog.my_schema.orders; for each table and record sizeInBytes, numFiles, createdAt and location.
3. Sum sizeInBytes and divide by 1073741824 for binary GB. Cloud storage totals may include older files outside the current snapshot.

Plan and admin role. Use a SQL warehouse or notebook with Unity Catalog access to the source tables. A deep clone also requires source read privileges, target table-creation privileges and appropriate external-location access; storage metrics require cloud-account access.

Exporting the full history

1. Arrange an external location pointing to a bucket or container you control.
2. Run CREATE TABLE my_catalog.export.orders DEEP CLONE my_catalog.my_schema.orders LOCATION 's3://my-export-bucket/orders'; for each table.
3. Copy the resulting external-location files using your cloud tools, such as aws s3 sync.

A deep clone copies the selected snapshot and metadata, not the source Delta version history or Unity Catalog tags.

Do not bypass Unity Catalog controls by directly reading managed-table paths; ordinary path access is unsupported outside Compatibility Mode.

Dropped managed tables have a limited recovery interval, seven days by default, during which UNDROP TABLE can recover them.

You don't need to run an export to talk to us, and there's no need to clean it first — removing personal data is part of our work.