This package allows you to run migrations on your PostgreSQL database using Golang Postgres client. See example for details.

You may also want to check go-pg-migrations before making a decision.

go-pg/migrations requires a Go version with Modules support and uses import path versioning. So please make sure to initialize a Go module:

go mod init github.com/my/repo
go get github.com/go-pg/migrations/v8

To run migrations on your project you should fulfill the following steps:

1. define the migration list;
2. implement an executable app that calls migration tool;
3. run migrations.

You can save SQL migration files at the same directory as your main.go file, they should have proper file extensions (more about migration files).

Migrations can be registered in the code using migrations.RegisterTx and migrations.MustRegisterTx functions. More details about migration registering.

You can run migrations from any place of your app or ecosystem. It can be a standalone application of a part of a big program, or maybe an HTTP handler, etc. Check example for some helpful information about practical usage.

Run migration tool by providing CLI arguments to the migrations.Run function.

Currently, the following arguments are supported:

• up - runs all available migrations;
• up [target] - runs available migrations up to the target one;
• down - reverts last migration;
• reset - reverts all migrations;
• version - prints current db version;
• set_version [version] - sets db version without running migrations.

You need to create database pg_migrations_example before running the example.

> cd example

> psql -c "CREATE DATABASE pg_migrations_example"
CREATE DATABASE

> go run *.go init
version is 0

> go run *.go version
version is 0

> go run *.go
creating table my_table...
adding id column...
seeding my_table...
migrated from version 0 to 4

> go run *.go version
version is 4

> go run *.go reset
truncating my_table...
dropping id column...
dropping table my_table...
migrated from version 4 to 0

> go run *.go up 2
creating table my_table...
adding id column...
migrated from version 0 to 2

> go run *.go
seeding my_table...
migrated from version 2 to 4

> go run *.go down
truncating my_table...
migrated from version 4 to 3

> go run *.go version
version is 3

> go run *.go set_version 1
migrated from version 3 to 1

> go run *.go create add email to users
created migration 5_add_email_to_users.go

Registers migrations to be executed inside transactions.

Registers migrations to be executed without any transaction.

SQL migrations are automatically picked up if placed in the same folder with main.go or Go migrations. SQL migrations may be manually registered using DiscoverSQLMigrations (from OS directory) or DiscoverSQLMigrationsFromFilesystem. It may be used with new go 1.16 embedding feature. Example:

//go:embed migrations/*.sql
var migrations embed.FS

collection := migrations.NewCollection()
collection.DiscoverSQLMigrationsFromFilesystem(http.FS(migrations), "migrations")

SQL migrations must have one of the following extensions:

• .up.sql - up migration;
• .down.sql - down migration;
• .tx.up.sql - transactional up migration;
• .tx.down.sql - transactional down migration.

By default SQL migrations are executed as single PostgreSQL statement. --gopg:split directive can be used to split migration into several statements:

SET statement_timeout = 60000;
SET lock_timeout = 60000;

--gopg:split

CREATE INDEX CONCURRENTLY ...;

By default, the migrations are executed outside without any transactions. Individual migrations can however be marked to be executed inside transactions by using the RegisterTx function instead of Register.

var oldVersion, newVersion int64

err := db.RunInTransaction(func(tx *pg.Tx) (err error) {
    oldVersion, newVersion, err = migrations.Run(tx, flag.Args()...)
    return
})
if err != nil {
    exitf(err.Error())
}