Rails Query - Group By with 2 groups

Question

In my website my users have an attribute localidade. This specifies where they live.

I'm trying to do a query where I group the results the following way:

localidade   | Number of Users
-------------+--------------
New York     | 6
Not New York | 8

I want the number of users from New York and the number of users from anywhere else but New York.

I tried this:

User.group("lower(localidade) = 'new york'").count

but since I don't have any users from new york and only 1 not from New York it returns:

{false => 1}

Am I able to give aliases to groups? Is there any way of grouping this way the results?

I'm gonna use the results for a Pie Graph from Graphkick.

Answer 1

You could write your query :

User.group("lower(localidade)")
    .select("CASE WHEN lower(localidade) = 'new york' THEN COUNT(id) END AS NewYork, 
             CASE WHEN lower(localidade) != 'new york' THEN COUNT(id) END AS Non-NewYork")

Since 9.4, you can use FILTER with aggregate expression :

User.group("lower(localidade)")
    .select("COUNT(id) FILTER (WHERE lower(localidade) != 'new york') AS NonNewyork, 
             COUNT(id) FILTER (WHERE lower(localidade) = 'new york') AS Newyork")

I created a Table to explain and test the above sql, and they worked as expected :

[shreyas@rails_app_test (master)]$ rails db
psql (9.4.1)
Type "help" for help.

app_development=# select id, location, name from people;
 id | location | name
----+----------+------
  2 | X        | foo
  3 | X        | foo
  4 | Y        | foo
(3 rows)

app_development=# SELECT COUNT(id) FILTER(WHERE lower(location) != 'x') AS Non_X_loc, COUNT(id) FILTER (WHERE lower(location) = 'x') AS X_loc FROM "people";
 non_x_loc | x_loc
-----------+-------
         1 |     2
(1 row)

Let me now, jump to the rails console, and test the equivalent Rails code :

[2] pry(main)> p = Person.select("COUNT(id) FILTER(WHERE lower(location) != 'x') AS Non_X_loc, COUNT(id) FILTER (WHERE lower(location) = 'x') AS X_loc ")
  Person Load (0.5ms)  SELECT COUNT(id) FILTER(WHERE lower(location) != 'x') AS Non_X_loc, COUNT(id) FILTER (WHERE lower(location) = 'x') AS X_loc  FROM "people"
=> [#<Person:0x007fd85ed71980 id: nil>]
[3] pry(main)> p.first.attributes
=> {"id"=>nil, "non_x_loc"=>1, "x_loc"=>2}
[6] pry(main)> Person.group("lower(location)").select("CASE WHEN lower(location) = 'x' THEN COUNT(id) END AS X_loc, CASE WHEN lower(location) != 'x' THEN COUNT(id) END AS Non_X_loc")
  Person Load (0.6ms)  SELECT CASE WHEN lower(location) = 'x' THEN COUNT(id) END AS X_loc, CASE WHEN lower(location) != 'x' THEN COUNT(id) END AS Non_X_loc FROM "people" GROUP BY lower(location)
=> [#<Person:0x007fd8608281e8 id: nil>, #<Person:0x007fd860828008 id: nil>]
[7] pry(main)> p = _
=> [#<Person:0x007fd8608281e8 id: nil>, #<Person:0x007fd860828008 id: nil>]
[8] pry(main)> p.map { |rec| rec.attributes }
=> [{"id"=>nil, "x_loc"=>nil, "non_x_loc"=>1}, {"id"=>nil, "x_loc"=>2, "non_x_loc"=>nil}]
[9] pry(main)> p.map { |rec| rec.attributes.except('id') }
=> [{"x_loc"=>nil, "non_x_loc"=>1}, {"x_loc"=>2, "non_x_loc"=>nil}]

Update

You can remove those nil from DB level only :

Rails code :

[shreyas@rails_app_test (master)]$ rails c
Loading development environment (Rails 4.2.0)
[1] pry(main)> Person.group("lower(location)").select("CASE WHEN lower(location) = 'x' THEN COUNT(id) ELSE 0 END AS X_loc, CASE WHEN lower(location) != 'x' THEN COUNT(id) ELSE 0 END AS Non_X_loc")
  Person Load (0.9ms)  SELECT CASE WHEN lower(location) = 'x' THEN COUNT(id) ELSE 0 END AS X_loc, CASE WHEN lower(location) != 'x' THEN COUNT(id) ELSE 0 END AS Non_X_loc FROM "people" GROUP BY lower(location)
=> [#<Person:0x007fd858c100b0 id: nil>, #<Person:0x007fd860853e88 id: nil>]
[2] pry(main)> p = _
=> [#<Person:0x007fd858c100b0 id: nil>, #<Person:0x007fd860853e88 id: nil>]
[3] pry(main)> p.map { |rec| rec.attributes }
=> [{"id"=>nil, "x_loc"=>0, "non_x_loc"=>1}, {"id"=>nil, "x_loc"=>2, "non_x_loc"=>0}]
[4] pry(main)> p.map { |rec| rec.attributes.except('id') }
=> [{"x_loc"=>0, "non_x_loc"=>1}, {"x_loc"=>2, "non_x_loc"=>0}]
[5] pry(main)> p = Person.select("count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, count(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc").group("lower(location)")
  Person Load (0.9ms)  SELECT count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, count(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc FROM "people" GROUP BY lower(location)
=> [#<Person:0x007fd85b150f78 id: nil>, #<Person:0x007fd85b150230 id: nil>]
[6] pry(main)> p.map { |rec| rec.attributes.except('id') }
=> [{"x_loc"=>0, "non_x_loc"=>1}, {"x_loc"=>2, "non_x_loc"=>0}]

SQL

app_development=# select CASE WHEN lower(location) = 'x' THEN COUNT(id) ELSE 0 END AS X_loc, CASE WHEN lower(location) != 'x' THEN COUNT(id) ELSE 0 END AS Non_X_loc from people group by lower(location);
 x_loc | non_x_loc
-------+-----------
     0 |         1
     2 |         0
(2 rows)
app_development=# select count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, count(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc from people group by lower(location);
 x_loc | non_x_loc
-------+-----------
     0 |         1
     2 |         0
(2 rows)

Update- II

The classical approach to get the output same as FILTER :

app_development=# select count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, sum(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc from people;
 x_loc | non_x_loc
-------+-----------
     2 |         1
(1 row)

app_development=# select sum(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, sum(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc from people;
 x_loc | non_x_loc
-------+-----------
     2 |         1
(1 row)

app_development=# select id, location, name from people;
 id | location | name
----+----------+------
  2 | X        | foo
  3 | X        | foo
  4 | Y        | foo
(3 rows)

app_development=#

And In Rails way :-

Loading development environment (Rails 4.2.0)
[1] pry(main)> p = Person.select("sum(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, sum(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc")
  Person Load (0.6ms)  SELECT sum(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, sum(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc FROM "people"
=> [#<Person:0x007fd85b6e6a78 id: nil>]
[2] pry(main)> p.first.attributes.except("id")
=> {"x_loc"=>2, "non_x_loc"=>1}
[3] pry(main)> p = Person.select("count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, count(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc")
  Person Load (0.5ms)  SELECT count(CASE WHEN lower(location) = 'x' THEN 1 END) AS X_loc, count(CASE WHEN lower(location) != 'x' THEN 1 END) AS Non_X_loc FROM "people"
=> [#<Person:0x007fd85b77f098 id: nil>]
[4] pry(main)> p.first.attributes.except("id")
=> {"x_loc"=>2, "non_x_loc"=>1}
[5] pry(main)>

Answer 2

Honestly, what you have works fine, you just need to understand that if there's no value in the hash for true (or for false for that matter) then the value must default to zero, you can do that with .to_i on what will be a nil value. So, eg.:

ny_count = User.group("lower(localidade) = 'new york'").count

"New York: #{ny_count[true].to_i}
Not New York: #{ny_count[false].to_i}
"