Concepts

This section covers three foundational Splitgraph concepts: repositories, images, and tags.

Familiar with Git?

If you know your way around git you already have a useful mental model. Working with Splitgraph repositories should feel quite similar to Git repositories in key ways. The only difference is that you're working with datasets vs code. It's possible to checkout, push, pull, and tag Splitgraph repos, just like git.

When you "checkout" a Splitgraph repo, you're checking out data into actual Postgres tables, instead of checking files out into the filesystem. These tables are available via your favorite Postgres client as you'd expect.

Also, similarly to how you can add your own remote to a git repo, it's possible to run your own Splitgraph engine and push up repos. (More at the decentralized docs.)

git command-line users might consider checking out the sgr command, which offers an intuitive way to work with Splitgraph (both cloud and self-hosted installs). The sgr docs cover what you can do.