What’s next
You have a project, an oplog (the file of schema changes) you can read, and the DDL it emits. These pages take it further.
Learn the model
- The oplog: what retrofit records, element identity, and replay.
- Sealing and deploying: sealed and unsealed ops, where each database is, and drift.
- Branches and merges: what a merge does to the oplog.
- Data files: where rows live beside the oplog.
- Principles: the rules retrofit upholds.
Ship a change
- Emit a migration script: the script that takes a database from one point in the oplog to another.
- Merge branches that change the oplog: two branches that each added changes.
- Check a pull request in CI: fail a pull request whose changes are risky to apply.
- Record a schema changed outside the oplog: catch the oplog up with a database someone altered by hand.
Shape the project
- Work with several schemas: register more schemas, scope them to directories, and pick one for a command.
- Exclude objects the project does not own: keep a provider’s tables, extensions and schemas out of the oplog.
- Work with row data: list the rows the emit commands apply, and write a record by hand.
Look a command up
Run any command with --help for its grammar and flags, or --help-all
for the extended vocabulary and caveats. The same text is on each
command’s page in the Command reference, and these pages cover the
rest:
- Output contract: what every command prints, and in what shape, for scripting against retrofit.
- Operators: every operator and the commands that handle it.
- Glossary: each term these pages use, in one sentence.
Compare it with your current tool
The demos on retrofit.sh show retrofit beside Flyway, Liquibase and Rails.