Schemify is a POC tool for applying schema declarations to a database. This is slightly different from applying migrations, like flyway, node-db-migrate, and most other schema migration tools. Migrations are effectively deltas, and in other systems all such deltas from the original schema are stored and applied.
This is a problem because it spreads out the declaration of the database in an unintuitive way.
A declaration is a description of the desired state of the database schema. Schemify compares the declarations against the database, builds the smallest migration to get from the current state to the desired state, and optionally applies those migrations.
Schemify is barely a proof of concept; it compares table declarations against an existing schema and can add tables or columns to tables. It doesn't yet understand simple things like indexes and certainly not complicated things like functions, custom types, or triggers. In principle these things are just extensions of what's already present, but it seems likely there will be some refactoring involved should we decide to move ahead with this.
Once this is getting built and deployed by a CI system, you should be able to install it via:
pip install schemify
You can install from source by running:
./setup pytest test
./setup install
Run schemify to install a given schema:
schemify -H {db-host} -p {db-port} -U {db-user} -P {db-password} -d {db-database} \
-v -s {schema-directory}
- CI needs to build and deploy this to dockerhub/artifactory.
- Need to support functions, custom types, triggers, and renaming of :allthethings:.
- A decision should be made whether the declarations should continue
to be yaml files, or whether an investment should be made in
parsing SQL so they can more naturally be written as
CREATE TABLE...statements. It's important to note that SQL doesn't have a native way to express that a schema entity used to have a different name.