Skip to content

Forking a Dataset

Forking creates a new dataset that starts with the same query definition (endpoint and filters) as an existing dataset. It is useful when you want to create a variation of a dataset — for example, querying a different subset of the same data — without modifying the original.

How to fork

  1. Navigate to Libraries → Datasets → Create New Dataset or go to /dataset/new
  2. Toggle Fork Existing Dataset on
  3. Select the Dataset Name you want to fork from
  4. Select the Version of that dataset to use as the starting point
  5. Give your new dataset a name
  6. Click Create Dataset

The new dataset is created with the same query (endpoint + filters) as the selected version of the source dataset. It gets a new datasetId and starts at version 1.

Forking vs. editing

Fork Edit existing
Creates a new dataset? Yes No
Affects the original? No Yes (new version)
Use when… You want a variation without changing the source You want to update the dataset in place

After forking

The Dataset Editor opens on the new dataset. You can immediately modify the filters, change the endpoint, or adjust deduplication settings without affecting the source dataset.