ActiveRecord::ConnectionAdapters::SchemaStatements

Adds a new check constraint to the table. expression is a String representation of verifiable boolean condition.

add_check_constraint :products, "price > 0", name: "price_check"

generates:

ALTER TABLE "products" ADD CONSTRAINT price_check CHECK (price > 0)

The options hash can include the following keys:

:name

The constraint name. Defaults to chk_rails_<identifier>.

:validate

(PostgreSQL only) Specify whether or not the constraint should be validated. Defaults to true.

Add a new type column named column_name to table_name.

The type parameter is normally one of the migrations native types, which is one of the following: :primary_key, :string, :text, :integer, :bigint, :float, :decimal, :numeric, :datetime, :time, :date, :binary, :boolean.

You may use a type not in this list as long as it is supported by your database (for example, “polygon” in MySQL), but this will not be database agnostic and should usually be avoided.

Available options are (none of these exists by default):

• :limit - Requests a maximum column length. This is the number of characters for a :string column and number of bytes for :text, :binary, and :integer columns. This option is ignored by some backends.

• :default - The column's default value. Use nil for NULL.

• :null - Allows or disallows NULL values in the column.

• :precision - Specifies the precision for the :decimal, :numeric, :datetime, and :time columns.

• :scale - Specifies the scale for the :decimal and :numeric columns.

• :collation - Specifies the collation for a :string or :text column. If not specified, the column will have the same collation as the table.

• :comment - Specifies the comment for the column. This option is ignored by some backends.

• :if_not_exists - Specifies if the column already exists to not try to re-add it. This will avoid duplicate column errors.

Note: The precision is the total number of significant digits, and the scale is the number of digits that can be stored following the decimal point. For example, the number 123.45 has a precision of 5 and a scale of 2. A decimal with a precision of 5 and a scale of 2 can range from -999.99 to 999.99.

Please be aware of different RDBMS implementations behavior with :decimal columns:

• The SQL standard says the default scale should be 0, :scale <= :precision, and makes no comments about the requirements of :precision.

• MySQL: :precision [1..63], :scale [0..30]. Default is (10,0).

• PostgreSQL: :precision [1..infinity], :scale [0..infinity]. No default.

• SQLite3: No restrictions on :precision and :scale, but the maximum supported :precision is 16. No default.

• Oracle: :precision [1..38], :scale [-84..127]. Default is (38,0).

• SqlServer: :precision [1..38], :scale [0..38]. Default (38,0).

Examples

add_column(:users, :picture, :binary, limit: 2.megabytes)
# ALTER TABLE "users" ADD "picture" blob(2097152)

add_column(:articles, :status, :string, limit: 20, default: 'draft', null: false)
# ALTER TABLE "articles" ADD "status" varchar(20) DEFAULT 'draft' NOT NULL

add_column(:answers, :bill_gates_money, :decimal, precision: 15, scale: 2)
# ALTER TABLE "answers" ADD "bill_gates_money" decimal(15,2)

add_column(:measurements, :sensor_reading, :decimal, precision: 30, scale: 20)
# ALTER TABLE "measurements" ADD "sensor_reading" decimal(30,20)

# While :scale defaults to zero on most databases, it
# probably wouldn't hurt to include it.
add_column(:measurements, :huge_integer, :decimal, precision: 30)
# ALTER TABLE "measurements" ADD "huge_integer" decimal(30)

# Defines a column that stores an array of a type.
add_column(:users, :skills, :text, array: true)
# ALTER TABLE "users" ADD "skills" text[]

# Defines a column with a database-specific type.
add_column(:shapes, :triangle, 'polygon')
# ALTER TABLE "shapes" ADD "triangle" polygon

# Ignores the method call if the column exists
add_column(:shapes, :triangle, 'polygon', if_not_exists: true)

Adds a new foreign key. from_table is the table with the key column, to_table contains the referenced primary key.

The foreign key will be named after the following pattern: fk_rails_<identifier>. identifier is a 10 character long string which is deterministically generated from the from_table and column. A custom name can be specified with the :name option.

Creating a simple foreign key

add_foreign_key :articles, :authors

generates:

ALTER TABLE "articles" ADD CONSTRAINT fk_rails_e74ce85cbc FOREIGN KEY ("author_id") REFERENCES "authors" ("id")

Creating a foreign key on a specific column

add_foreign_key :articles, :users, column: :author_id, primary_key: "lng_id"

generates:

ALTER TABLE "articles" ADD CONSTRAINT fk_rails_58ca3d3a82 FOREIGN KEY ("author_id") REFERENCES "users" ("lng_id")

Creating a cascading foreign key

add_foreign_key :articles, :authors, on_delete: :cascade

generates:

ALTER TABLE "articles" ADD CONSTRAINT fk_rails_e74ce85cbc FOREIGN KEY ("author_id") REFERENCES "authors" ("id") ON DELETE CASCADE

The options hash can include the following keys:

:column

The foreign key column name on from_table. Defaults to to_table.singularize + "_id"

:primary_key

The primary key column name on to_table. Defaults to id.

:name

The constraint name. Defaults to fk_rails_<identifier>.

:on_delete

Action that happens ON DELETE. Valid values are :nullify, :cascade and :restrict

:on_update

Action that happens ON UPDATE. Valid values are :nullify, :cascade and :restrict

:validate

(PostgreSQL only) Specify whether or not the constraint should be validated. Defaults to true.

Adds a new index to the table. column_name can be a single Symbol, or an Array of Symbols.

The index will be named after the table and the column name(s), unless you pass :name as an option.

Creating a simple index

add_index(:suppliers, :name)

generates:

CREATE INDEX index_suppliers_on_name ON suppliers(name)

Creating a index which already exists

add_index(:suppliers, :name, if_not_exists: true)

generates:

CREATE INDEX IF NOT EXISTS index_suppliers_on_name ON suppliers(name)

Note: Not supported by MySQL.

Creating a unique index

add_index(:accounts, [:branch_id, :party_id], unique: true)

generates:

CREATE UNIQUE INDEX index_accounts_on_branch_id_and_party_id ON accounts(branch_id, party_id)

Creating a named index

add_index(:accounts, [:branch_id, :party_id], unique: true, name: 'by_branch_party')

generates:

CREATE UNIQUE INDEX by_branch_party ON accounts(branch_id, party_id)

Creating an index with specific key length

add_index(:accounts, :name, name: 'by_name', length: 10)

generates:

CREATE INDEX by_name ON accounts(name(10))

Creating an index with specific key lengths for multiple keys

add_index(:accounts, [:name, :surname], name: 'by_name_surname', length: {name: 10, surname: 15})

generates:

CREATE INDEX by_name_surname ON accounts(name(10), surname(15))

Note: SQLite doesn't support index length.

Creating an index with a sort order (desc or asc, asc is the default)

add_index(:accounts, [:branch_id, :party_id, :surname], name: 'by_branch_desc_party', order: {branch_id: :desc, party_id: :asc})

generates:

CREATE INDEX by_branch_desc_party ON accounts(branch_id DESC, party_id ASC, surname)

Note: MySQL only supports index order from 8.0.1 onwards (earlier versions accepted the syntax but ignored it).

Creating a partial index

add_index(:accounts, [:branch_id, :party_id], unique: true, where: "active")

generates:

CREATE UNIQUE INDEX index_accounts_on_branch_id_and_party_id ON accounts(branch_id, party_id) WHERE active

Note: Partial indexes are only supported for PostgreSQL and SQLite.

Creating an index with a specific method

add_index(:developers, :name, using: 'btree')

generates:

CREATE INDEX index_developers_on_name ON developers USING btree (name) -- PostgreSQL
CREATE INDEX index_developers_on_name USING btree ON developers (name) -- MySQL

Note: only supported by PostgreSQL and MySQL

Creating an index with a specific operator class

add_index(:developers, :name, using: 'gist', opclass: :gist_trgm_ops)
# CREATE INDEX developers_on_name ON developers USING gist (name gist_trgm_ops) -- PostgreSQL

add_index(:developers, [:name, :city], using: 'gist', opclass: { city: :gist_trgm_ops })
# CREATE INDEX developers_on_name_and_city ON developers USING gist (name, city gist_trgm_ops) -- PostgreSQL

add_index(:developers, [:name, :city], using: 'gist', opclass: :gist_trgm_ops)
# CREATE INDEX developers_on_name_and_city ON developers USING gist (name gist_trgm_ops, city gist_trgm_ops) -- PostgreSQL

Note: only supported by PostgreSQL

Creating an index with a specific type

add_index(:developers, :name, type: :fulltext)

generates:

CREATE FULLTEXT INDEX index_developers_on_name ON developers (name) -- MySQL

Note: only supported by MySQL.

Creating an index with a specific algorithm

add_index(:developers, :name, algorithm: :concurrently)
# CREATE INDEX CONCURRENTLY developers_on_name on developers (name)

Note: only supported by PostgreSQL.

Concurrently adding an index is not supported in a transaction.

For more information see the “Transactional Migrations” section.

Adds a reference. The reference column is a bigint by default, the :type option can be used to specify a different type. Optionally adds a _type column, if :polymorphic option is provided. add_reference and add_belongs_to are acceptable.

The options hash can include the following keys:

:type

The reference column type. Defaults to :bigint.

:index

Add an appropriate index. Defaults to true. See add_index for usage of this option.

:foreign_key

Add an appropriate foreign key constraint. Defaults to false, pass true to add. In case the join table can't be inferred from the association pass :to_table with the appropriate table name.

:polymorphic

Whether an additional _type column should be added. Defaults to false.

:null

Whether the column allows nulls. Defaults to true.

Create a user_id bigint column without an index

add_reference(:products, :user, index: false)

Create a user_id string column

add_reference(:products, :user, type: :string)

Create supplier_id, supplier_type columns

add_reference(:products, :supplier, polymorphic: true)

Create a supplier_id column with a unique index

add_reference(:products, :supplier, index: { unique: true })

Create a supplier_id column with a named index

add_reference(:products, :supplier, index: { name: "my_supplier_index" })

Create a supplier_id column and appropriate foreign key

add_reference(:products, :supplier, foreign_key: true)

Create a supplier_id column and a foreign key to the firms table

add_reference(:products, :supplier, foreign_key: { to_table: :firms })

Adds timestamps (created_at and updated_at) columns to table_name. Additional options (like :null) are forwarded to add_column.

add_timestamps(:suppliers, null: true)

Changes the column's definition according to the new options. See TableDefinition#column for details of the options you can use.

change_column(:suppliers, :name, :string, limit: 80)
change_column(:accounts, :description, :text)

Changes the comment for a column or removes it if nil.

Passing a hash containing :from and :to will make this change reversible in migration:

change_column_comment(:posts, :state, from: "old_comment", to: "new_comment")

Sets a new default value for a column:

change_column_default(:suppliers, :qualification, 'new')
change_column_default(:accounts, :authorized, 1)

Setting the default to nil effectively drops the default:

change_column_default(:users, :email, nil)

Passing a hash containing :from and :to will make this change reversible in migration:

change_column_default(:posts, :state, from: nil, to: "draft")

Sets or removes a NOT NULL constraint on a column. The null flag indicates whether the value can be NULL. For example

change_column_null(:users, :nickname, false)

says nicknames cannot be NULL (adds the constraint), whereas

change_column_null(:users, :nickname, true)

allows them to be NULL (drops the constraint).

The method accepts an optional fourth argument to replace existing NULLs with some other value. Use that one when enabling the constraint if needed, since otherwise those rows would not be valid.

Please note the fourth argument does not set a column's default.

A block for changing columns in table.

# change_table() yields a Table instance
change_table(:suppliers) do |t|
  t.column :name, :string, limit: 60
  # Other column alterations here
end

The options hash can include the following keys:

:bulk

Set this to true to make this a bulk alter query, such as

ALTER TABLE `users` ADD COLUMN age INT, ADD COLUMN birthdate DATETIME ...

Defaults to false.

Only supported on the MySQL and PostgreSQL adapter, ignored elsewhere.

Add a column

change_table(:suppliers) do |t|
  t.column :name, :string, limit: 60
end

Change type of a column

change_table(:suppliers) do |t|
  t.change :metadata, :json
end

Add 2 integer columns

change_table(:suppliers) do |t|
  t.integer :width, :height, null: false, default: 0
end

Add created_at/updated_at columns

change_table(:suppliers) do |t|
  t.timestamps
end

Add a foreign key column

change_table(:suppliers) do |t|
  t.references :company
end

Creates a company_id(bigint) column.

Add a polymorphic foreign key column

change_table(:suppliers) do |t|
  t.belongs_to :company, polymorphic: true
end

Creates company_type(varchar) and company_id(bigint) columns.

Remove a column

change_table(:suppliers) do |t|
  t.remove :company
end

Remove several columns

change_table(:suppliers) do |t|
  t.remove :company_id
  t.remove :width, :height
end

Remove an index

change_table(:suppliers) do |t|
  t.remove_index :company_id
end

See also Table for details on all of the various column transformations.

Changes the comment for a table or removes it if nil.

Passing a hash containing :from and :to will make this change reversible in migration:

change_table_comment(:posts, from: "old_comment", to: "new_comment")

Returns an array of check constraints for the given table. The check constraints are represented as CheckConstraintDefinition objects.

Checks to see if a column exists in a given table.

# Check a column exists
column_exists?(:suppliers, :name)

# Check a column exists of a particular type
column_exists?(:suppliers, :name, :string)

# Check a column exists with a specific definition
column_exists?(:suppliers, :name, :string, limit: 100)
column_exists?(:suppliers, :name, :string, default: 'default')
column_exists?(:suppliers, :name, :string, null: false)
column_exists?(:suppliers, :tax, :decimal, precision: 8, scale: 2)

Returns an array of Column objects for the table specified by table_name.

Creates a new join table with the name created using the lexical order of the first two arguments. These arguments can be a String or a Symbol.

# Creates a table called 'assemblies_parts' with no id.
create_join_table(:assemblies, :parts)

You can pass an options hash which can include the following keys:

:table_name

Sets the table name, overriding the default.

:column_options

Any extra options you want appended to the columns definition.

:options

Any extra options you want appended to the table definition.

:temporary

Make a temporary table.

:force

Set to true to drop the table before creating it. Defaults to false.

Note that create_join_table does not create any indices by default; you can use its block form to do so yourself:

create_join_table :products, :categories do |t|
  t.index :product_id
  t.index :category_id
end

Add a backend specific option to the generated SQL (MySQL)

create_join_table(:assemblies, :parts, options: 'ENGINE=InnoDB DEFAULT CHARSET=utf8')

generates:

CREATE TABLE assemblies_parts (
  assembly_id bigint NOT NULL,
  part_id bigint NOT NULL,
) ENGINE=InnoDB DEFAULT CHARSET=utf8

Creates a new table with the name table_name. table_name may either be a String or a Symbol.

There are two ways to work with create_table. You can use the block form or the regular form, like this:

Block form

# create_table() passes a TableDefinition object to the block.
# This form will not only create the table, but also columns for the
# table.

create_table(:suppliers) do |t|
  t.column :name, :string, limit: 60
  # Other fields here
end

Block form, with shorthand

# You can also use the column types as method calls, rather than calling the column method.
create_table(:suppliers) do |t|
  t.string :name, limit: 60
  # Other fields here
end

Regular form

# Creates a table called 'suppliers' with no columns.
create_table(:suppliers)
# Add a column to 'suppliers'.
add_column(:suppliers, :name, :string, {limit: 60})

The options hash can include the following keys:

:id

Whether to automatically add a primary key column. Defaults to true. Join tables for ActiveRecord::Base.has_and_belongs_to_many should set it to false.

A Symbol can be used to specify the type of the generated primary key column.

:primary_key

The name of the primary key, if one is to be added automatically. Defaults to id. If :id is false, then this option is ignored.

If an array is passed, a composite primary key will be created.

Note that Active Record models will automatically detect their primary key. This can be avoided by using self.primary_key= on the model to define the key explicitly.

:options

Any extra options you want appended to the table definition.

:temporary

Make a temporary table.

:force

Set to true to drop the table before creating it. Set to :cascade to drop dependent objects as well. Defaults to false.

:if_not_exists

Set to true to avoid raising an error when the table already exists. Defaults to false.

:as

SQL to use to generate the table. When this option is used, the block is ignored, as are the :id and :primary_key options.

Add a backend specific option to the generated SQL (MySQL)

create_table(:suppliers, options: 'ENGINE=InnoDB DEFAULT CHARSET=utf8mb4')

generates:

CREATE TABLE suppliers (
  id bigint auto_increment PRIMARY KEY
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4

Rename the primary key column

create_table(:objects, primary_key: 'guid') do |t|
  t.column :name, :string, limit: 80
end

generates:

CREATE TABLE objects (
  guid bigint auto_increment PRIMARY KEY,
  name varchar(80)
)

Change the primary key column type

create_table(:tags, id: :string) do |t|
  t.column :label, :string
end

generates:

CREATE TABLE tags (
  id varchar PRIMARY KEY,
  label varchar
)

Create a composite primary key

create_table(:orders, primary_key: [:product_id, :client_id]) do |t|
  t.belongs_to :product
  t.belongs_to :client
end

generates:

CREATE TABLE order (
    product_id bigint NOT NULL,
    client_id bigint NOT NULL
);

ALTER TABLE ONLY "orders"
  ADD CONSTRAINT orders_pkey PRIMARY KEY (product_id, client_id);

Do not add a primary key column

create_table(:categories_suppliers, id: false) do |t|
  t.column :category_id, :bigint
  t.column :supplier_id, :bigint
end

generates:

CREATE TABLE categories_suppliers (
  category_id bigint,
  supplier_id bigint
)

Create a temporary table based on a query

create_table(:long_query, temporary: true,
  as: "SELECT * FROM orders INNER JOIN line_items ON order_id=orders.id")

generates:

CREATE TEMPORARY TABLE long_query AS
  SELECT * FROM orders INNER JOIN line_items ON order_id=orders.id

See also TableDefinition#column for details on how to create columns.

Checks to see if the data source name exists on the database.

data_source_exists?(:ebooks)

Returns the relation names useable to back Active Record models. For most adapters this means all tables and views.

Drops the join table specified by the given arguments. See create_join_table for details.

Although this command ignores the block if one is given, it can be helpful to provide one in a migration's change method so it can be reverted. In that case, the block will be used by create_join_table.

Drops a table from the database.

:force

Set to :cascade to drop dependent objects as well. Defaults to false.

:if_exists

Set to true to only drop the table if it exists. Defaults to false.

Although this command ignores most options and the block if one is given, it can be helpful to provide these in a migration's change method so it can be reverted. In that case, options and the block will be used by create_table.

Checks to see if a foreign key exists on a table for a given foreign key definition.

# Checks to see if a foreign key exists.
foreign_key_exists?(:accounts, :branches)

# Checks to see if a foreign key on a specified column exists.
foreign_key_exists?(:accounts, column: :owner_id)

# Checks to see if a foreign key with a custom name exists.
foreign_key_exists?(:accounts, name: "special_fk_name")

Returns an array of foreign keys for the given table. The foreign keys are represented as ForeignKeyDefinition objects.

Checks to see if an index exists on a table for a given index definition.

# Check an index exists
index_exists?(:suppliers, :company_id)

# Check an index on multiple columns exists
index_exists?(:suppliers, [:company_id, :company_type])

# Check a unique index exists
index_exists?(:suppliers, :company_id, unique: true)

# Check an index with a custom name exists
index_exists?(:suppliers, :company_id, name: "idx_company_id")

Verifies the existence of an index with a given name.

Returns an array of indexes for the given table.

Returns a hash of mappings from the abstract data types to the native database types. See TableDefinition#column for details on the recognized abstract data types.

Returns just a table's primary key

Removes the given check constraint from the table.

remove_check_constraint :products, name: "price_check"

The expression parameter will be ignored if present. It can be helpful to provide this in a migration's change method so it can be reverted. In that case, expression will be used by add_check_constraint.

Removes the column from the table definition.

remove_column(:suppliers, :qualification)

The type and options parameters will be ignored if present. It can be helpful to provide these in a migration's change method so it can be reverted. In that case, type and options will be used by add_column. Indexes on the column are automatically removed.

If the options provided include an if_exists key, it will be used to check if the column does not exist. This will silently ignore the migration rather than raising if the column was already used.

remove_column(:suppliers, :qualification, if_exists: true)

Removes the given columns from the table definition.

remove_columns(:suppliers, :qualification, :experience)

type and other column options can be passed to make migration reversible.

remove_columns(:suppliers, :qualification, :experience, type: :string, null: false)

Removes the given foreign key from the table. Any option parameters provided will be used to re-add the foreign key in case of a migration rollback. It is recommended that you provide any options used when creating the foreign key so that the migration can be reverted properly.

Removes the foreign key on accounts.branch_id.

remove_foreign_key :accounts, :branches

Removes the foreign key on accounts.owner_id.

remove_foreign_key :accounts, column: :owner_id

Removes the foreign key on accounts.owner_id.

remove_foreign_key :accounts, to_table: :owners

Removes the foreign key named special_fk_name on the accounts table.

remove_foreign_key :accounts, name: :special_fk_name

The options hash accepts the same keys as SchemaStatements#add_foreign_key with an addition of

:to_table

The name of the table that contains the referenced primary key.

Removes the given index from the table.

Removes the index on branch_id in the accounts table if exactly one such index exists.

remove_index :accounts, :branch_id

Removes the index on branch_id in the accounts table if exactly one such index exists.

remove_index :accounts, column: :branch_id

Removes the index on branch_id and party_id in the accounts table if exactly one such index exists.

remove_index :accounts, column: [:branch_id, :party_id]

Removes the index named by_branch_party in the accounts table.

remove_index :accounts, name: :by_branch_party

Removes the index on branch_id named by_branch_party in the accounts table.

remove_index :accounts, :branch_id, name: :by_branch_party

Checks if the index exists before trying to remove it. Will silently ignore indexes that don't exist.

remove_index :accounts, if_exists: true

Removes the index named by_branch_party in the accounts table concurrently.

remove_index :accounts, name: :by_branch_party, algorithm: :concurrently

Note: only supported by PostgreSQL.

Concurrently removing an index is not supported in a transaction.

For more information see the “Transactional Migrations” section.

Removes the reference(s). Also removes a type column if one exists. remove_reference and remove_belongs_to are acceptable.

Remove the reference

remove_reference(:products, :user, index: false)

Remove polymorphic reference

remove_reference(:products, :supplier, polymorphic: true)

Remove the reference with a foreign key

remove_reference(:products, :user, foreign_key: true)

Removes the timestamp columns (created_at and updated_at) from the table definition.

remove_timestamps(:suppliers)

Renames a column.

rename_column(:suppliers, :description, :name)

Renames an index.

Rename the index_people_on_last_name index to index_users_on_last_name:

rename_index :people, 'index_people_on_last_name', 'index_users_on_last_name'

Renames a table.

rename_table('octopuses', 'octopi')

Truncates a table alias according to the limits of the current adapter.

Returns the table comment that's stored in database metadata.

Checks to see if the table table_name exists on the database.

table_exists?(:developers)

Returns an array of table names defined in the database.

Checks to see if the view view_name exists on the database.

view_exists?(:ebooks)

Returns an array of view names defined in the database.