Multiple aggregate queries using the sequel gem - ruby

Question

Is it possible to use sequel to execute a query like this:
select (select count(*) from users where blah = 'blah') as "users",
(select count(*) from contacts where blah = 'blah') as "contacts"
I know I can execute these queries one a time using sequel, but I would like to execute them all at the same time.

Answer 1

You can do that query without writing raw SQL with the following:
dataset = DB.select {[
DB[:users].where(blah: 'blah').select { count('*') }.as(:users),
DB[:contacts].where(blah: 'blah').select { count('*') }.as(:contacts)
]}
dataset.first
# => { users: X, contacts: Y }
dataset.sql
# => "SELECT (SELECT count('*') FROM \"users\" WHERE (\"blah\" = 'blah')) AS \"users\",
# (SELECT count('*') FROM \"contacts\" WHERE (\"blah\" = 'blah')) AS \"contacts\""

Answer 2

Yes you can do it ok with the sequel gem.
require 'sequel'
DB = Sequel.sqlite # memory database
DB.create_table :users do
primary_key :id
String :name
end
users = DB[:users] # Create a dataset
users.insert(:name => 'jim')
DB.create_table :contacts do
primary_key :id
String :name
end
contacts = DB[:contacts] # Create a dataset
contacts.insert(:name => 'first')
DB['select (select count(*) from users where name = "jim") as users,
(select count(*) from contacts where name = "first") as contacts'].each do |row|
puts row
end
#{:users=>1, :contacts=>1}
It should be noted however that it is not a great idea to include raw strings in a sequel method.
They can be extracted as follows:
DB['select (select count(*) from users where name = ?) as users,
(select count(*) from contacts where name = ?) as contacts, 'jim', 'first'].each do |row|
puts row
end
Also as noted in another answer you could express this query completely without resorting to SQL which is more in the spirit of the module. :
dataset = DB.select {[
DB[:users].where(name: 'jim').select { count('*') }.as(:users),
DB[:contacts].where(name: 'first').select { count('*') }.as(:contacts)
]}
dataset.sql
# SELECT (SELECT count('*') FROM `users` WHERE (`name` = 'jim')) AS 'users', (SELECT count('*') FROM `contacts` WHERE (`name` = 'first')) AS 'contacts'
dataset.first
# {:users=>1, :contacts=>1}

Answer 3

You can execute as you mentioned above, there is no issue with your query .
I have executed the same and it was executed without any issues ..
below was what i have executed
select( SELECT count(*) FROM `abcde` where `User_ID`=4001) as "aa",
(SELECT count(*) FROM `abcdef` where `User_ID`=4018) as "bbb"
results :
aa bbb
181 364

Answer 4

you need to add * from after select and may be some parenthesis:
select * from ((select count(*) from users where blah = 'blah') as "users",
(select count(*) from contacts where blah = 'blah') as "contacts")