Creating a PostgreSQL sequence to a field (which is not the ID of the record)

Question

I am working on a Ruby on Rails app. We are using a PostgreSQL database.

There is a table named scores with the following columns:

Column        | Type
--------------+-----------------------
id            | integer
value         | double precision
ran_at        | timestamp
active        | boolean
build_id      | bigint
metric_id     | integer
platform_id   | integer
mode_id       | integer
machine_id    | integer
higher_better | boolean
job_id        | integer
variation_id  | integer
step          | character varying(255)

I need to add a sequence to job_id (note: there is no model for job).

How do I create this sequence?

Answer 1

So I figured out how to do this using ActiveRecord migrations on Ruby on Rails. I basically used Erwin's commands and help from this page and put them in the migration files. These are the steps:

1. In the terminal, type:

rails g migration CreateJobIdSequence
rails g migration AddJobIdSequenceToScores

2. Edit the migration files as follows:

20140709181616_create_job_id_sequence.rb :

class CreateJobIdSequence < ActiveRecord::Migration
  def up
    execute <<-SQL
      CREATE SEQUENCE job_id_seq;
    SQL
  end

  def down
    execute <<-SQL
      DROP SEQUENCE job_id_seq;
    SQL
  end
end

20140709182313_add_job_id_sequence_to_scores.rb :

class AddJobIdSequenceToScores < ActiveRecord::Migration
  def up
    execute <<-SQL
      ALTER SEQUENCE job_id_seq OWNED BY scores.job_id;
      ALTER TABLE scores ALTER COLUMN job_id SET DEFAULT nextval('job_id_seq');
    SQL
  end

  def down
    execute <<-SQL
      ALTER SEQUENCE job_id_seq OWNED BY NONE;
      ALTER TABLE scores ALTER COLUMN job_id SET NOT NULL;
    SQL
  end
end

3. Migrate the database. In the terminal type:

rake db:migrate

Answer 2

Use CREATE SEQUENCE:

CREATE SEQUENCE scores_job_id_seq;  -- = default name for plain a serial

Then add a column default to scores.job_id:

ALTER TABLE scores ALTER COLUMN job_id SET DEFAULT nextval('scores_job_id_seq');

If you want to bind the sequence to the column (so it is deleted when the column is deleted), also run:

ALTER SEQUENCE scores_job_id_seq OWNED BY scores.job_id;

All of this can be replaced with using the pseudo data type serial for the column job_id to begin with:

• Safely and cleanly rename tables that use serial primary key columns in Postgres?

If your table already has rows, you may want to set the SEQUENCE to the next highest value and fill in missing serial values in the table:

SELECT setval('scores_job_id_seq', COALESCE(max(job_id), 1)) FROM scores;

Optionally:

UPDATE scores
SET    job_id = nextval('scores_job_id_seq')
WHERE  job_id IS NULL;

• How to check a sequence efficiently for used and unused values in PostgreSQL
• Postgres manually alter sequence
• How to reset postgres' primary key sequence when it falls out of sync?

The only remaining difference, a serial column is also set to NOT NULL. You may or may not want that, too:

ALTER TABLE scores ALTER COLUMN job_id SET NOT NULL;

But you cannot just alter the type of an existing integer:

ALTER TABLE scores ALTER job_id TYPE serial;

serial is not an actual data type. It's just a notational convenience feature for CREATE TABLE.
In Postgres 10 or later consider an IDENTITY column:

• Auto increment table column