All data uploaded to a table is versioned. All tables in a dataset are versioned together. A version is a locked copy of the data that can't be edited or deleted. This means that you can confidently use it in your projects, knowing that it will never change while you're working on it.

If the dataset's owner has an update, finds a mistake to correct, or would otherwise like to modify the existing data they can create and release a new version that will exist alongside the previous version of that dataset.

Every version contains table(s) of data along with the edits applied to modify it from the previous version, and any release notes written by the editor.

Version history

You can see what version you're on by looking in the top right corner of a dataset page or node. To view a previous version, click on the version dropdown on the top right of the dataset page or node.

You can view the the version history and release notes of a dataset on the dataset page overview.

Within each version you can view the Edits made to create it if you have the appropriate access level.

As versions age, their corresponding tables may become archived. You won't be able to view the cells on archived tables, but you can still query them on the query tab or work with them in a project.

Compare versions

On the dataset page you can choose to compare the version you are looking at to a previous version of this same dataset. Any changes will be shown in red/green highlights on the page.

Semantic version tags

To help researchers better understand if they should upgrade to a new version, Redivis uses semantic versioning, of the form v[major].[minor] The first version of every dataset is v1.0. For subsequent versions, the tag will augment automatically depending on the changes being released.

  • Major update: Your existing code may not run.

    • Triggered when any edits rename, delete, or retype variables

  • Minor update: Your existing code will generally run.

    • Edits include adding / removing records, recoding variables, adding variables