You can store SQL queries in views that you can version in Git. In "enterprise" databases like Postgres, Oracle and MSSQL, you can also store statements in proc

You can store SQL queries in views that you can version in Git. In "enterprise" databases like Postgres, Oracle and MSSQL, you can also store statements in procedures/functions that you can also store in Git. We also do Git based migration scripts that create new / update existing views, tables, procedures, data, etc. This happens automatically on test servers and can be rolled out to production. It is somewhat tricky to build such an infrastructure and differs from "traditional" backend software, but it is what it is with Databases that heavily rely on state. The benefit is that you save a lot of bandwidth when you process the data on the machine where they live, and the architecture of the IT landscape is way simpler.