Counting by earliest date found from an inner join? - postgresql

Question

I have two tables, customerusermap and users. Whenever a user signs up with our product, they immediately get added into a table called users but it isn't until they start paying for a user that they get added to a table called customerusermap.
The users table looks like this:
id | customer_id | firstname | lastname | created_at
-------------------------------------------------------
1725 | cus_3hEmhErE2jbwsO | Abby | Smith | 2015-03-19
1726 | cus_7oNweUrE4jbwr2 | Sam | Peters | 2015-06-20
The customerusermap table looks like this:
customer_id | user_id | created_at
------------------------------------------
cus_3hEmhErE2jbwsO | 9275 | 2015-09-01
cus_3hEmhErE2jbwsO | 2628 | 2015-09-05
cus_3hEmhErE2jbwsO | 2358 | 2015-07-05
cus_3hEmhErE2jbwsO | 3158 | 2015-08-05
cus_3hEmhErE2jbwsO | 2487 | 2015-08-05
cus_3hEmhErE2jbwsO | 6044 | 2015-08-05
cus_7oNweUrE4jbwr2 | 8094 | 2015-08-25
cus_7oNweUrE4jbwr2 | 2345 | 2015-09-02
In this example, Abby(cus_3hEmhErE2jbwsO) is paying for 6 users. She started paying for user 2358 2015-07-05 so she should be considered a paying customer 07-2015, not 03-2015. Sam is paying for 2 users and he started paying for user 8094 in 08-2015 so he is considered to be a paying customer for 08-2015, not 06-2015. I have a query that grabs and groups by the number of paying customers each month:
SELECT concat(extract(MONTH from u.created_at),'-',extract(year from u.created_at)) as "Month",
COUNT(distinct u.email) as "Total AB Paying Customers"
FROM customerusermap AS cm, users AS u
WHERE cm.customer_id=u.customer_id AND cm.user_id <> u.id
GROUP BY 1,extract(month from u.created_at),extract(year from u.created_at)
ORDER BY extract(year from u.created_at),extract(month from u.created_at);
But this grabs and counts by the date the customer was added to the users table, not the date they actually started paying. How would I grab the counts so that it grabs for the earliest date in the customerusermap table? What the needed output should look like in this example is:
Month | Total AB Paying Customers
-------------------------------------
07-2015 | 1
08-2015 | 1

Answer 1

You can use the following query:
SELECT CONCAT(EXTRACT(MONTH FROM startedPayingDate), '-',
EXTRACT(YEAR FROM startedPayingDate)) AS "Month",
COUNT(*) AS "Total AB Paying Customers"
FROM (
SELECT customer_id, MIN(created_at) AS startedPayingDate
FROM customerusermap AS cm
WHERE NOT EXISTS (SELECT 1
FROM users AS u
WHERE cm.user_id = u.id)
GROUP BY customer_id ) AS t
GROUP BY 1
I used a NOT EXISTS operator to exclude records that relate to 'paying for themselves' customers (if that is really your intention).
Once you get the MIN(created_at) date per customer_id, then you can easily count per date in an outer query.
Demo here