Class: Sequel::Schema::AlterTableGenerator

Inherits:
Object
  • Object
show all
Defined in:
lib/sequel/database/schema_generator.rb

Overview

Schema::AlterTableGenerator is an internal class that the user is not expected to instantiate directly. Instances are created by Database#alter_table. It is used to specify table alteration parameters. It takes a Database object and a block of operations to perform on the table, and gives the Database an array of table altering operations, which the database uses to alter a table’s description.

For more information on Sequel’s support for schema modification, see the “Schema Modification” guide.

Direct Known Subclasses

Postgres::AlterTableGenerator

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(db, &block) ⇒ AlterTableGenerator

Set the Database object to which to apply the changes, and evaluate the block in the context of this object.



343
344
345
346
347
 
# File 'lib/sequel/database/schema_generator.rb', line 343

def initialize(db, &block)
  @db = db
  @operations = []
  instance_exec(&block) if block
end

Instance Attribute Details

#operations ⇒ Object (readonly)

An array of operations to perform



339
340
341
 
# File 'lib/sequel/database/schema_generator.rb', line 339

def operations
  @operations
end

Instance Method Details

#add_column(name, type, opts = OPTS) ⇒ Object

Add a column with the given name, type, and opts. See CreateTableGenerator#column for the available options.

add_column(:name, String) # ADD COLUMN name varchar(255)

PostgreSQL specific options:

:if_not_exists

Set to true to not add the column if it already exists (PostgreSQL 9.6+)

MySQL specific options:

:after

The name of an existing column that the new column should be positioned after

:first

Create this new column before all other existing columns



362
363
364
365
 
# File 'lib/sequel/database/schema_generator.rb', line 362

def add_column(name, type, opts = OPTS)
  @operations << {:op => :add_column, :name => name, :type => type}.merge!(opts)
  nil
end

#add_constraint(name, *args, &block) ⇒ Object

Add a constraint with the given name and args. See CreateTableGenerator#constraint.

add_constraint(:valid_name, Sequel.like(:name, 'A%'))
# ADD CONSTRAINT valid_name CHECK (name LIKE 'A%' ESCAPE '\')
add_constraint({name: :valid_name, deferrable: true}, Sequel.like(:name, 'A%'))
# ADD CONSTRAINT valid_name CHECK (name LIKE 'A%' ESCAPE '\') DEFERRABLE INITIALLY DEFERRED



374
375
376
377
378
 
# File 'lib/sequel/database/schema_generator.rb', line 374

def add_constraint(name, *args, &block)
  opts = name.is_a?(Hash) ? name : {:name=>name}
  @operations << opts.merge(:op=>:add_constraint, :type=>:check, :check=>block || args)
  nil
end

#add_foreign_key(name, table, opts = OPTS) ⇒ Object

Add a foreign key with the given name and referencing the given table. See CreateTableGenerator#column for the available options.

You can also pass an array of column names for creating composite foreign keys. In this case, it will assume the columns exist and will only add the constraint. You can provide a :name option to name the constraint.

NOTE: If you need to add a foreign key constraint to a single existing column use the composite key syntax even if it is only one column.

add_foreign_key(:artist_id, :table) # ADD COLUMN artist_id integer REFERENCES table
add_foreign_key([:name], :table) # ADD FOREIGN KEY (name) REFERENCES table

PostgreSQL specific options:

:not_valid

Set to true to add the constraint with the NOT VALID syntax. This makes it so that future inserts must respect referential integrity, but allows the constraint to be added even if existing column values reference rows that do not exist. After all the existing data has been cleaned up, validate_constraint can be used to mark the constraint as valid. Note that this option only makes sense when using an array of columns.



413
414
415
416
 
# File 'lib/sequel/database/schema_generator.rb', line 413

def add_foreign_key(name, table, opts = OPTS)
  return add_composite_foreign_key(name, table, opts) if name.is_a?(Array)
  add_column(name, Integer, {:table=>table}.merge!(opts))
end

#add_full_text_index(columns, opts = OPTS) ⇒ Object

Add a full text index on the given columns. See CreateTableGenerator#index for available options.



420
421
422
 
# File 'lib/sequel/database/schema_generator.rb', line 420

def add_full_text_index(columns, opts = OPTS)
  add_index(columns, {:type=>:full_text}.merge!(opts))
end

#add_index(columns, opts = OPTS) ⇒ Object

Add an index on the given columns. See CreateTableGenerator#index for available options.

add_index(:artist_id) # CREATE INDEX table_artist_id_index ON table (artist_id)

Options:

:name

Give a specific name for the index. Highly recommended if you plan on dropping the index later.

:where

A filter expression, used to setup a partial index (if supported).

:unique

Create a unique index.

PostgreSQL specific options:

:concurrently

Create the index concurrently, so it doesn’t require an exclusive lock on the table.

:index_type

The underlying index type to use for a full_text index, gin by default).

:language

The language to use for a full text index (simple by default).

:opclass

Set an opclass to use for all columns (per-column opclasses require custom SQL).

:type

Set the index type (e.g. full_text, spatial, hash, gin, gist, btree).

:if_not_exists

Only create the index if an index of the same name doesn’t already exists

MySQL specific options:

:type

Set the index type, with full_text and spatial indexes handled specially.

Microsoft SQL Server specific options:

:include

Includes additional columns in the index.

:key_index

Sets the KEY INDEX to the given value.

:type

clustered uses a clustered index, full_text uses a full text index.



456
457
458
459
 
# File 'lib/sequel/database/schema_generator.rb', line 456

def add_index(columns, opts = OPTS)
  @operations << {:op => :add_index, :columns => Array(columns)}.merge!(opts)
  nil
end

#add_primary_key(name, opts = OPTS) ⇒ Object

Add a primary key. See CreateTableGenerator#column for the available options. Like add_foreign_key, if you specify the column name as an array, it just creates a constraint:

add_primary_key(:id) # ADD COLUMN id serial PRIMARY KEY
add_primary_key([:artist_id, :name]) # ADD PRIMARY KEY (artist_id, name)



467
468
469
470
471
 
# File 'lib/sequel/database/schema_generator.rb', line 467

def add_primary_key(name, opts = OPTS)
  return add_composite_primary_key(name, opts) if name.is_a?(Array)
  opts = @db.serial_primary_key_options.merge(opts)
  add_column(name, opts.delete(:type), opts)
end

#add_spatial_index(columns, opts = OPTS) ⇒ Object

Add a spatial index on the given columns. See CreateTableGenerator#index for available options.



475
476
477
 
# File 'lib/sequel/database/schema_generator.rb', line 475

def add_spatial_index(columns, opts = OPTS)
  add_index(columns, {:type=>:spatial}.merge!(opts))
end

#add_unique_constraint(columns, opts = OPTS) ⇒ Object

Add a unique constraint to the given column(s)

add_unique_constraint(:name) # ADD UNIQUE (name)
add_unique_constraint(:name, name: :unique_name) # ADD CONSTRAINT unique_name UNIQUE (name)

Supports the same :deferrable option as CreateTableGenerator#column.



386
387
388
389
 
# File 'lib/sequel/database/schema_generator.rb', line 386

def add_unique_constraint(columns, opts = OPTS)
  @operations << {:op => :add_constraint, :type => :unique, :columns => Array(columns)}.merge!(opts)
  nil
end

#drop_column(name, opts = OPTS) ⇒ Object

Remove a column from the table.

drop_column(:artist_id) # DROP COLUMN artist_id
drop_column(:artist_id, cascade: true) # DROP COLUMN artist_id CASCADE

Options:

:cascade

CASCADE the operation, dropping other objects that depend on the dropped column.

PostgreSQL specific options:

:if_exists

Use IF EXISTS, so no error is raised if the column does not exist.



492
493
494
495
 
# File 'lib/sequel/database/schema_generator.rb', line 492

def drop_column(name, opts=OPTS)
  @operations << {:op => :drop_column, :name => name}.merge!(opts)
  nil
end

#drop_constraint(name, opts = OPTS) ⇒ Object

Remove a constraint from the table:

drop_constraint(:unique_name) # DROP CONSTRAINT unique_name
drop_constraint(:unique_name, cascade: true) # DROP CONSTRAINT unique_name CASCADE

MySQL/SQLite specific options:

:type

Set the type of constraint to drop, either :primary_key, :foreign_key, or :unique.



506
507
508
509
 
# File 'lib/sequel/database/schema_generator.rb', line 506

def drop_constraint(name, opts=OPTS)
  @operations << {:op => :drop_constraint, :name => name}.merge!(opts)
  nil
end

#drop_foreign_key(name, opts = OPTS) ⇒ Object

Remove a foreign key and the associated column from the table. General options:

:name

The name of the constraint to drop. If not given, uses the same name that would be used by add_foreign_key with the same columns.

NOTE: If you want to drop only the foreign key constraint but keep the column, use the composite key syntax even if it is only one column.

drop_foreign_key(:artist_id) # DROP CONSTRAINT table_artist_id_fkey, DROP COLUMN artist_id
drop_foreign_key([:name]) # DROP CONSTRAINT table_name_fkey



521
522
523
524
525
526
527
528
 
# File 'lib/sequel/database/schema_generator.rb', line 521

def drop_foreign_key(name, opts=OPTS)
  if !name.is_a?(Array) && opts[:foreign_key_constraint_name]
    opts = Hash[opts]
    opts[:name] = opts[:foreign_key_constraint_name]
  end
  drop_composite_foreign_key(Array(name), opts)
  drop_column(name) unless name.is_a?(Array)
end

#drop_index(columns, options = OPTS) ⇒ Object

Remove an index from the table. General options:

:name

The name of the index to drop. If not given, uses the same name that would be used by add_index with the same columns.

PostgreSQL specific options:

:cascade

Cascade the index drop to dependent objects.

:concurrently

Drop the index using CONCURRENTLY, which doesn’t block operations on the table. Supported in PostgreSQL 9.2+.

:if_exists

Only drop the index if it already exists.

drop_index(:artist_id) # DROP INDEX table_artist_id_index
drop_index([:a, :b]) # DROP INDEX table_a_b_index
drop_index([:a, :b], name: :foo) # DROP INDEX foo



545
546
547
548
 
# File 'lib/sequel/database/schema_generator.rb', line 545

def drop_index(columns, options=OPTS)
  @operations << {:op => :drop_index, :columns => Array(columns)}.merge!(options)
  nil
end

#rename_column(name, new_name, opts = OPTS) ⇒ Object

Rename one of the table’s columns.

rename_column(:name, :artist_name) # RENAME COLUMN name TO artist_name



553
554
555
556
 
# File 'lib/sequel/database/schema_generator.rb', line 553

def rename_column(name, new_name, opts = OPTS)
  @operations << {:op => :rename_column, :name => name, :new_name => new_name}.merge!(opts)
  nil
end

#set_column_allow_null(name, allow_null = true) ⇒ Object

Set a given column as allowing NULL values.

set_column_allow_null(:artist_name) # ALTER COLUMN artist_name DROP NOT NULL

On MySQL, make sure to use a symbol for the name of the column, as otherwise you can lose the default and type for the column.



594
595
596
597
 
# File 'lib/sequel/database/schema_generator.rb', line 594

def set_column_allow_null(name, allow_null=true)
  @operations << {:op => :set_column_null, :name => name, :null => allow_null}
  nil
end

#set_column_default(name, default) ⇒ Object

Modify the default value for one of the table’s column.

set_column_default(:artist_name, 'a') # ALTER COLUMN artist_name SET DEFAULT 'a'

To remove an existing default value, use nil as the value:

set_column_default(:artist_name, nil) # ALTER COLUMN artist_name SET DEFAULT NULL

On MySQL, make sure to use a symbol for the name of the column, as otherwise you can lose the type and NULL/NOT NULL setting for the column.



568
569
570
571
 
# File 'lib/sequel/database/schema_generator.rb', line 568

def set_column_default(name, default)
  @operations << {:op => :set_column_default, :name => name, :default => default}
  nil
end

#set_column_not_null(name) ⇒ Object

Set a given column as not allowing NULL values.

set_column_not_null(:artist_name) # ALTER COLUMN artist_name SET NOT NULL

On MySQL, make sure to use a symbol for the name of the column, as otherwise you can lose the default and type for the column.



605
606
607
 
# File 'lib/sequel/database/schema_generator.rb', line 605

def set_column_not_null(name)
  set_column_allow_null(name, false)
end

#set_column_type(name, type, opts = OPTS) ⇒ Object

Modify the type of one of the table’s column.

set_column_type(:artist_name, 'char(10)') # ALTER COLUMN artist_name TYPE char(10)

PostgreSQL specific options:

:using

Add a USING clause that specifies how to convert existing values to new values.

On MySQL, make sure to use a symbol for the name of the column, as otherwise you can lose the default and NULL/NOT NULL setting for the column.



583
584
585
586
 
# File 'lib/sequel/database/schema_generator.rb', line 583

def set_column_type(name, type, opts=OPTS)
  @operations << {:op => :set_column_type, :name => name, :type => type}.merge!(opts)
  nil
end