This is a Database Migrations system for CakePHP 3.0.
The plugin consists of a wrapper for the phinx migrations library.
Full documentation of the plugin can be found on the CakePHP Cookbook.
You can install this plugin into your CakePHP application using composer.
Run the following command
composer require cakephp/migrations
You will need to add the following line to your application's bootstrap.php file:
Plugin::load('Migrations');
Additionally, you will need to configure the default
database configuration in your config/app.php
file.
After creating/modifying a migration file, you can run your changes in the database by executing:
# The following will run migrations against the default database connection
bin/cake migrations migrate
# Rolling back migrations. If a `change()` method is defined, it will be reversed.
# Otherwise, the `down()` method will be invoked
bin/cake migrations rollback
# Retrieve the status of executed and pending migrations
bin/cake migrations status
# All console commands can take a `--plugin` or `-p` option
bin/cake migrations status -p PluginName
# You can also scope a command to a connection via the `--connection` or `-c` option
bin/cake migrations status -c my_datasource
# The following will mark migrations as marked without actually running it.
bin/cake migrations mark_migrated
# DEPRECATED: The use of the argument `all` will have the same effect as above
bin/cake migrations mark_migrated all
# Using the option `--target` it will try to mark every migration from beginning up to the given VERSION
bin/cake migrations mark_migrated --target=VERSION
# When using the `--target` option you can also use `--exclude` or `--only`:
# `--exclude` will try to mark every migration from beginning until the given VERSION (excluding it from marking)
# `--only` will try to mark only the given VERSION
bin/cake migrations mark_migrated --target=VERSION --exclude
bin/cake migrations mark_migrated --target=VERSION --only
# DEPRECATED: Using the VERSION argument will try to mark only the given VERSION
bin/cake migrations mark_migrated VERSION
This plugin provides two interfaces to creating migrations: a passthru to the Phinx library and a way to use the bake
utility.
The Phinx Migrations shell can be invoked via the following command from your application's root folder:
$ bin/cake migrations
The command above will display a list of available options. Make sure you read the official phinx documentation to understand how migrations are created and executed in your database.
Please note that you need to learn how to write your own migrations.
Empty migrations files will be created leaving you to fill in the up() and down() or change() if you want automatically reversible migrations.
Once again, please make sure you read the official phinx documentation to understand how migrations are created and executed in your database.
You can also use the bake
command to generate migrations.
# The following will create an initial snapshot migration file:
bin/cake bake migration_snapshot Initial
# Create an empty migration file
bin/cake bake migration AddFieldToTable
# You can specify a plugin to bake into
bin/cake bake migration AddFieldToTable --plugin PluginName
# You can specify an alternative connection when generating a migration.
bin/cake bake migration AddFieldToTable --connection connection
# Require that the table class exists before creating a migration
bin/cake bake migration AddFieldToTable --require-table
These commands will create a file under config/Migrations
with the current
database snapshot as the contents of the change()
method. You may edit this
as desired.
Please note that you will need to learn how to write your own migrations, you need to fill in the up() and down() or change() methods if you want automatically reversible migrations.
Once again, please make sure you read the official phinx documentation to understand how migrations are created and executed in your database.
To create a table called statuses
and use a CHAR (36) for the id
field, this requires you to turn off the id.
See:
$table = $this->table('statuses',
[
'id' => false,
'primary_key' => ['id']
]);
$table->addColumn('id', 'char', ['limit' => 36])
->addColumn('name', 'char', ['limit' => 255])
->addColumn('model', 'string', ['limit' => 128])
->create();
Phinx automatically creates an auto-increment
id
field for every table. This will hopefully be fixed in the future.
If you need to create a table with a different collation than the database default one, you can define it
with the table
method, as an option :
$table = $this
->table('categories', [
'collation' => 'latin1_german1_ci'
])
->addColumn('title', 'string', [
'default' => null,
'limit' => 255,
'null' => false,
])
->create();
Note however this can only be done on table creation : there is currently no way of adding a column to an existing table with a different collation than the table or the database. Only MySQL and SqlServer supports this configuration key for the time being.
When using this option, you can still modify the migration before running them if so desired.
You can optionally generate entire migration files from the CLI without
interacting with the database or an editor. This functionality only works when
arguments are passed to the command bin/cake bake generate
as follows:
bin/cake bake migration create_users name:string created modified
bin/cake bake migration alter_users name:string:index
bin/cake bake migration drop_users
bin/cake bake migration add_taxonomic_stuff_to_posts category:string tags:string
bin/cake bake migration remove_taxonomic_stuff_from_posts category tags
The above commands would:
- Create a users table with the fields [
id
,name
,created
,modified
]. A single primary key index would exist onid
- as phinx autogenerates the field and it's index - and thecreated
andmodified
fields would default todatetime
, as per CakePHP conventions. Since the type is specified onname
, it is string. - Add an index to the
name
column in theusers
table. - Drop the users table.
- Add
category
andtags
fields to theposts
table. - Remove
category
andtags
fields from theposts
table.
Due to the conventions, not all schema changes can be performed via these shell commands.
Migration Names can follow any of the following patterns:
- create_table
/^(Create)(.*)/
: Creates the specified table - drop_table
/^(Drop)(.*)/
: Drops the specified table. Ignores specified field arguments. - add_field
/^(Add).*(?:To)(.*)/
: Adds fields to the specified table - remove_field
/^(Remove).*(?:From)(.*)/
: Removes fields from the specified table - alter_table
/^(Alter)(.*)/
: Alters the specified table. The alter_table command can be used as an alias forCreateTable
andAddField
.
Migration names are used as migration class names, and thus may collide with other migrations if the class names are not unique. In this case, it may be necessary to manually override the name at a later date, or simply change the name you are specifying.
Fields are verified via the following the following regular expression:
/^(\w*)(?::(\w*\[?\d*\]?))?(?::(\w*))?(?::(\w*))?/
They follow the format:
field:fieldType[length]:indexType:indexName
The length parameter for the fieldType
is optional and should always be
written between bracket.
For instance, the following are all valid ways of specifying the primary key id
:
id:primary_key
id:primary_key:primary
id:integer:primary
id:integer:primary:ID_INDEX
Field types are those generically made available by phinx.
There are some heuristics to choosing fieldtypes when left unspecified or set to an invalid value:
id
: integercreated
,modified
,updated
: datetime- Default string
You can specify the wanted length for a field type by writing it between bracket:
username:string[128]
If no length is specified, lengths for certain columns are defaulted:
- string:
255
- integer:
11
- biginteger:
20