I have table "orders" with column ID (id order) and IDU (id user).
I want calculate like this:
50 person have 1 order
30 person have 2 order
SELECT DISTNICT count(id), count(idu) FROM orders GROUP BY idu
I call this a histogram of histogram queries. You want to know how many users have a given order count. The solution is to use subqueries and
group by twice:
select cnt, count(*), min(idu), max(idu) from (select idu, count(*) as cnt from orders group by idu ) ou group by cnt;
When running this type of query, I usually include the min and max of the user id, so I can readily find examples of users (typically those that have many orders).