Skip to content

Backport a dataset to ETL

This tutorial might be partial or incomplete. Please check with the team if you have questions.

Before ETL came along, datasets were uploaded directly to our Grapher admin site. If you want to leverage these datasets in the ETL, they won't be available out of the box, since they weren't initially imported via ETL. You have two main options to bring them into ETL:

  1. Migrate them into the ETL pipeline.
  2. Move them to Fast-Track.

Fast-Track is recommended for users who prefer working on Google sheets rather than the ETL environment. However, it's not as flexible as ETL, and it's not recommended for datasets that require frequent updates.

How to Migrate to ETL

Automatically generate ETL steps for a specific dataset with the following command. Choose the appropriate namespace, version and short name for the migrated dataset. etl-backport-migrate --dataset-id your_dataset_id --namespace your_namespace --version your_version --short-name your_short_name

After running this, follow the terminal prompts.

How to Migrate to Fast-Track

To move a dataset to Fast-Track, generate a Google spreadsheet using: backport-fasttrack --dataset-id xxx --short-name your_short_name

Once generated, open the Fast-Track spreadsheet and follow the provided instructions.

Using automatically backported datasets via backport://backport/owid/latest/dataset_xxx_... is discouraged.