A Leiningen plugin to manage SQL database migrations simply and dynamically with clojure/jdbc.
Note: Apologies for this repository going offline (deleted my github during the Microsoft acquisition), but this project is active and will be maintained. Thanks very much for using / supporting it!
- Database agnostic (migrations are created from sequences of sql strings)
- Supports creation of stored procedures / other complicated sql statements
- Supports running migrations under multiple environments with "ENV" environment variable
- Runs within clojure so sql strings can easily be dynamically constructed
- Has a simple
create
command for generating migration files
Put clj-sql-up into your plugins vector:
:plugins [[clj-sql-up "0.4.2"]]
Add database connection info (both your driver and jdbc connection string):
:clj-sql-up {:database "jdbc:postgresql://foo@127.0.0.1:5432/foo"
:deps [[org.postgresql/postgresql "42.2.23"]]}
;; OR (might be stale; haven't used mysql driver in a while)
:clj-sql-up {:database {:subprotocol "mysql"
:subname "//127.0.0.1:3306/foo"
:user "foo"
:password ""}
:deps [[mysql/mysql-connector-java "8.0.26"]]}
;; OR (for use with multiple environments or a custom repo dependency)
:clj-sql-up {:database-test "jdbc:postgresql://foo@127.0.0.1:5432/foo_test"
:deps [[org.postgresql/postgresql "42.2.23"]]
;; Note, this is included already in your migrations, but
;; is an example of adding additional dependency repos
:repos { "clojars" "http://clojars.org/repo" }}
Basic usage (though it doesn't get much more complicated):
$ lein clj-sql-up create create-posts
;; migrations/20130714150641624-create-posts.clj
;; Note: the generated methods would just return a blank vector
;; Just inserting some statements as an example
(defn up []
["CREATE TABLE foo(id int)"
"CREATE INDEX foo_index ON foo(id)"])
(defn down []
["drop table foo"])
$ lein clj-sql-up migrate
Migrating: 20130714150634587-create-timestamps-fn.clj
Migrating: 20130714150641624-create-posts.clj
$ lein clj-sql-up rollback
Reversing: 20130714150641624-create-posts.clj
To run migrations on a different database, ensure you define the connection in your project.clj file (see examples above) and run:
$ ENV=test lein clj-sql-up migrate
$ ENV=test lein clj-sql-up rollback
And this will run the migration on the database specfied by :database-test in your config section
Heroku uses leiningen 1.x by default which introduces some dependency conflicts for pomegranate (used by clj-sql-up). See this issue for more detail: https://github.com/ckuttruff/clj-sql-up/issues/5
Please include the following within your project.clj to use leiningen 2.x in your app: :min-lein-version "2.0.0"
- Add more tests
- Make certain aspects more generic (specifying migrations dir, etc.)
The motivation behind clj-sql-up is to create a migration library that is as simple as possible, but allows for migration statements to exist within the context of clojure (not as plain sql files). It also allows for creation of more complicated sql commands (eg: stored procedures/functions) - something certain clojure migration libraries currently fail to do properly.
- @peterschwarz
- added jdbc 0.3.3 compatibility
- created additional tests with hsqldb
- ensured db-agnostic migration table creation
- @kitallis
- Fixed issue with use of java's Long casting
- @adomokos
- Updates to dynamic binding of migrations directory
- @icambron
- Updates to create.clj to respect rebinding of migrations dir
- Additional testing around migration file creation
- @cdorrat
- Allow migrations to be loaded from classpath
- Permit packaging of migrations with a jar/uberjar
Copyright © 2018 Christopher Kuttruff
Distributed under the Eclipse Public License, the same as Clojure.