Data source guides

How to export your Databricks data

Last checked 28 September 2026

Answer

SQL writes the data out in bulk: INSERT OVERWRITE [LOCAL] DIRECTORY directory_path USING file_format [OPTIONS (...)] query, where file_format is TEXT, CSV, JSON, JDBC, PARQUET, ORC, HIVE or LIBSVM, and the directory can be a Unity Catalog volume path such as /Volumes/my_catalog/my_schema/my_volume/. LOCAL targets a local file system directory instead. For a result set read in the browser, click the downward arrow beside the results tab title and pick CSV, or Excel where the notebook runs on a SQL warehouse or serverless compute; the file saves locally under the notebook name. Downloading results is on by default and a workspace admin can switch it off in the workspace settings.

What it holds

Delta tables in Unity Catalog with their rows, plus notebooks and the query results run against them.

Find the size

Databricks shows no single storage figure for a workspace. Run DESCRIBE DETAIL on a Delta table, which returns sizeInBytes, the size of the latest snapshot of the table, and numFiles, the number of files in the latest version. Do not read the count off the screen: notebook table results are limited to 10,000 rows or 2 MB, whichever is lower, and job clusters cap notebook output at 30 MB.

What to enter on the form

Databricks in GB, and the year your company started using it. Seller form

Sources