Suppose I created a table table in a Rails app. Some time later, I add a column running:
rails generate migration AddUser_idColumnToTable user_id:string.
Then I realize I need to add user_id as an index. I know about the add_index method, but where should this method be called? Am I supposed to run a migration (if yes, which one ?), then adding by hand this method?
You can run another migration, just for the index:
class AddIndexToTable < ActiveRecord::Migration
def change
add_index :table, :user_id
end
end
:table is the actual table name, so in the case of a users table, you'd substitute :users for :table.
add_index :bars, :foo where the table name is barsIf you need to create a user_id then it would be a reasonable assumption that you are referencing a user table. In which case the migration shall be:
rails generate migration AddUserRefToProducts user:references
This command will generate the following migration:
class AddUserRefToProducts < ActiveRecord::Migration
def change
add_reference :user, :product, index: true
end
end
After running rake db:migrate both a user_id column and an index will be added to the products table.
In case you just need to add an index to an existing column, e.g. name of a user table, the following technique may be helpful:
rails generate migration AddIndexToUsers name:string:index will generate the following migration:
class AddIndexToUsers < ActiveRecord::Migration
def change
add_column :users, :name, :string
add_index :users, :name
end
end
Delete add_column line and run the migration.
In the case described you could have issued rails generate migration AddIndexIdToTable index_id:integer:index command and then delete add_column line from the generated migration. But I'd rather recommended to undo the initial migration and add reference instead:
rails generate migration RemoveUserIdFromProducts user_id:integer
rails generate migration AddUserRefToProducts user:references
user_id column was already in production, then removing it will delete real data, so that's a bad idea. If it is still in development, why not just revert and then edit the original migration? Or if you feel the need to leave the original migration code untouched, you could add comments mentioning the relationship between the two migrations.
rails generate migration RemoveUserIdFromProducts user_id:integerwill look like?Add in the generated migration after creating the column the following (example)
add_index :photographers, :email, :unique => true
For references you can call
rails generate migration AddUserIdColumnToTable user:references
If in the future you need to add a general index you can launch this
rails g migration AddOrdinationNumberToTable ordination_number:integer:index
Generated code:
class AddOrdinationNumberToTable < ActiveRecord::Migration
def change
add_column :tables, :ordination_number, :integer
add_index :tables, :ordination_number, unique: true
end
end
For those who are using postgresql db and facing error
StandardError: An error has occurred, this and all later migrations canceled:
=== Dangerous operation detected #strong_migrations ===
Adding an index non-concurrently blocks writes
please refer this article
example:
class AddAncestryToWasteCodes < ActiveRecord::Migration[6.0]
disable_ddl_transaction!
def change
add_column :waste_codes, :ancestry, :string
add_index :waste_codes, :ancestry, algorithm: :concurrently
end
end
You can use this, just think Job is the name of the model to which you are adding index cader_id:
class AddCaderIdToJob < ActiveRecord::Migration[5.2]
def change
change_table :jobs do |t|
t.integer :cader_id
t.index :cader_id
end
end
end
