Skip to content

Latest commit

 

History

History
227 lines (166 loc) · 8.7 KB

README.md

File metadata and controls

227 lines (166 loc) · 8.7 KB

PGX Schema - Embedded Database Migration Library for Go (PGX Driver Version)

An embeddable library for tracking and applying modifications to the PostgreSQL schema from inside a Go application using the jackc/pgx driver.

go.dev reference CircleCI Build Status Go Report Card Code Coverage

NOTE: If you use a database/sql driver instead, please see the related adlio/schema package.

Features

  • Cloud-friendly design tolerates embedded use in clusters
  • Supports migrations in embed.FS (requires go:embed in Go 1.16+)
  • Depends only on Go standard library and jackc/pgx (Note that all go.mod dependencies are used only in tests)
  • Unidirectional migrations (no "down" migration complexity)

Usage Instructions

Create a pgxschema.Migrator in your bootstrap/config/database connection code, then call its Apply() method with your database connection and a slice of *pgxschema.Migration structs.

The .Apply() function figures out which of the supplied Migrations have not yet been executed in the database (based on the ID), and executes the Script for each in alphabetical order by ID.

The []*pgxschema.Migration can be created manually, but the package has some utility functions to make it easier to read .sql files into structs, with the filename as the ID and the contents being the Script.

Using go:embed (requires Go 1.16+)

Go 1.16 added features to embed a directory of files into the binary as an embedded filesystem (embed.FS).

Assuming you have a directory of SQL files called my-migrations/ next to your main.go file, you'll run something like this (the comments with go:embed are relevant).

//go:embed my-migrations
var MyMigrations embed.FS

func main() {
   db, err := pgxpool.Connect() // or pgx.Connect()

   migrator := pgxschema.NewMigrator()
   err = migrator.Apply(
      db,
      pgxschema.FSMigrations(MyMigrations, "my-migrations/*.sql"),
   )
}

The result will be a slice of *pgxschema.Migration{} with the file's name (without the extension) as the ID property, and the entire contents of the file as its Script property. The test-migrations/saas directory provides an example.

Using Inline Migration Structs

If you're running an earlier version of Go, Migration{} structs will need to be created manually:

db, err := pgxpool.Connect() // or pgx.Connect()

migrator := pgxschema.NewMigrator()
migrator.Apply(db, []*pgxschema.Migration{
   &pgxschema.Migration{
      ID: "2019-09-24 Create Albums",
      Script: `
      CREATE TABLE albums (
         id SERIAL PRIMARY KEY,
         title CHARACTER VARYING (255) NOT NULL
      )
      `,
   },
})

Constructor Options

The NewMigrator() function accepts option arguments to customize its behavior.

WithTableName

By default, the tracking table will be placed in the schema from the search_path, and it will be named schema_migrations. This behavior can be changed by supplying a WithTableName() option to the NewMigrator() call.

m := pgxschema.NewMigrator(pgxschema.WithTableName("my_migrations"))

If you need to customize both the schema and the table name, provide two arguments:

m := pgxschema.NewMigrator(pgxschema.WithTableName("my_schema", "my_migrations"))

NOTE: Providing a schema like so does not influence the behavior of SQL run inside your migrations. If a migration needs to CREATE TABLE in a specific schema, that will need to be specified inside the migration itself or configured via the search_path when opening a connection.

It is theoretically possible to create multiple Migrators and to use mutliple migration tracking tables within the same application and database.

Concurrent Execution Support

The pgxschema package utilizes PostgreSQL Advisory Locks to ensure that only one process can run migrations at a time.

This allows multiple processes (not just goroutines) to run Apply on identically configured migrators simultaneously. The first-arriving process will win and perform all needed migrations on the database. All other processses will wait until the lock is released, after which they'll each obtain the lock and run Apply() which should be a no-op based on the first-arriving process' successful completion.

Migration Ordering

Migrations are not executed in the order they are specified in the slice. They will be re-sorted alphabetically by their IDs before executing them.

Rules for Writing Migrations

  1. Never, ever change the ID (filename) or Script (file contents) of a Migration which has already been executed on your database. If you've made a mistake, you'll need to correct it in a subsequent migration.

  2. Use a consistent, but descriptive format for migration IDs/filenames. Consider prefixing them with today's timestamp. Examples:

      ID: "2019-01-01T13:45 Creates Users"
      ID: "2019-01-10T10:33 Creates Artists"
    

    Do not use simple sequential numbers like ID: "1" with a distributed team unless you have a reliable process for developers to "claim" the next ID.

Contributions

... are welcome. Please include tests with your contribution. We've integrated dockertest to automate the process of creating clean test databases.

Testing Procedure

Testing requires a Docker daemon running on your test machine to spin-up temporary PostgreSQL database servers to test against. Ensure your contribution keeps test coverage high and passes all existing tests.

go test -v -cover

Package Opinions

There are many other schema migration tools. This one exists because of a particular set of opinions:

  1. Database credentials are runtime configuration details, but database schema is a build-time applicaton dependency, which means it should be "compiled in" to the build, and should not rely on external tools.
  2. Using an external command-line tool for schema migrations needlessly complicates testing and deployment.
  3. SQL is the best language to use to specify changes to SQL schemas.
  4. "Down" migrations add needless complication, aren't often used, and are tedious to properly test when they are used. In the unlikely event you need to migrate backwards, it's possible to write the "rollback" migration as a separate "up" migration.
  5. Deep dependency chains should be avoided, especially in a compiled binary. We don't want to import an ORM into our binaries just to get SQL querying support. The pgxschema package imports only standard library packages and the jackc/pgx driver code. (NOTE *We do import ory/dockertest to automate testing on various PostgreSQL versions via docker).

Roadmap

  • Port adlio/schema to a jackc/pgx-friendly version
  • Alter transaction handling to be more PostgreSQL-specific
  • 100% test coverage, including running against multiple PostgreSQL versions
  • Support for creating []*Migration from a Go 1.16 embed.FS
  • Documentation for using Go 1.16 // go:embed to populate Script variables
  • Options for alternative failure behavior when pg_advisory_lock() takes too long. The current behavior should allow for early failure by providing a context with a timeout to WithContext(), but this hasn't been tested.
  • Add a Validate() method to allow checking migration names for consistency and to detect problematic changes in the migrations list

Version History

1.0.1 - Jun 19, 2022

  • Update downstream dependencies of ory/dockertest due to security issues.

1.0.0 - Jan 4, 2022

  • Add support for migrations in an embed.FS (FSMigrations(filesystem fs.FS, glob string))
  • Update go.mod to go 1.17
  • Simplify Apply() routine, improve test coverage
  • Security updates to upstream dependencies

0.0.3 - Dec 10, 2021

Security updates to upstream dependencies.

0.0.2 - Nov 18, 2021

Security updates to upstream dependencies.

0.0.1 - Oct 7, 2021

First port from adlio/schema.

License

Copyright (c) 2022 Aaron Longwell, released under the MIT License.