具有多个条件和子选择的SQL查询

Question

I am fairly new to using joins and selecting from multiple tables. I have successfully figured out how to get data from 2 tables based on a number of conditions. Now my issue is that I need to return a count only if another condition is also met. The idea is like this: I have a query that returns the number of items in a specific status at a specific time. My query for this is:

SELECT count(*),
       e.campus_id,
       e.course_id
FROM statuses_history AS sh,
     enrolments AS e
WHERE sh.date_added > '2015-08-01 00:00:00'
  AND sh.date_added < '2015-08-20 23:59:59'
  AND sh.status_id = 57
  AND sh.item_id = e.enrolment_id
  AND (e.course_id = 2
       OR e.course_id =7
       OR e.course_id = 8
       OR e.course_id = 9)
GROUP BY e.campus_id,
         e.course_id;

Now I have to check that it has been in a different status, lets say sh.status_id = 50 before it was in 57. It also does not have to fall within the data range specified. So I basically need to change my query in some way to also select if it's ever been in status_id 50 and then only return the result if both the statuses are found. sh.status_id =57 will be bound by the dates specified.

Thanks in advance for the help.

Answer 1

You can adjust the subselect as needed, but this will return a row if another value is found in the database.

I don't like subselects, but this should work. I have not tested the syntax.

select count(*), e.campus_id, e.course_id
FROM statuses_history as sh, enrolments as e
WHERE sh.date_added > '2015-08-01 00:00:00'
      AND sh.date_added < '2015-08-20 23:59:59'
      AND sh.status_id = 57
      AND sh.item_id = e.enrolment_id
      AND (e.course_id = 2 OR e.course_id =7
          OR e.course_id = 8 OR e.course_id = 9)
      AND EXISTS (select sh2.status_id from statuses_history as sh2
                 WHERE NOT sh2.status_id = 57
                   AND sh2.item_id = e.enrolment_id)
GROUP BY e.campus_id, e.course_id;

Answer 2

I assume that the primary key is item_id?

You just have to join the same tables by the primary key and the status_id = 50.

For this case I will use INNER JOIN (my preference) but you can use FROM/WHERE CLAUSE.

select count(*), e.campus_id, e.course_id  
FROM statuses_history as sh, enrolments as e

INNER JOIN statuses_history AS oldSh ON oldSh.item_id = sh.item_id AND oldSh.status_id = 50

WHERE sh.date_added > '2015-08-01 00:00:00' AND sh.date_added < '2015-08-20 23:59:59' AND sh.status_id = 57 and sh.item_id = e.enrolment_id AND (e.course_id = 2 OR e.course_id =7 OR e.course_id = 8 OR e.course_id = 9) 
GROUP BY e.campus_id, e.course_id;

But if you prefer WHERE, you can do the following:

select count(*), e.campus_id, e.course_id  
FROM statuses_history as sh, enrolments as e, statuses_history AS oldSh ON  
WHERE oldSh.item_id = sh.item_id AND oldSh.status_id = 50 AND 
sh.date_added > '2015-08-01 00:00:00' AND sh.date_added < '2015-08-20 23:59:59' AND sh.status_id = 57 and sh.item_id = e.enrolment_id AND (e.course_id = 2 OR e.course_id =7 OR e.course_id = 8 OR e.course_id = 9) 
GROUP BY e.campus_id, e.course_id;

Answer 3

Add a left join and use a case statement to do your count.

LEFT JOIN statuses_history sh2 ON sh2.item_id = e.enrollment_id AND sh2.status_id = 50

You'll need to use a different format for your inner joins in order to do the left join. Here's the new query:

select count(sh.status_id), e.campus_id, e.course_id  , SUM(CASE WHEN sh2.status_id IS NOT NULL THEN 1 ELSE 0 END) as cnt
FROM statuses_history sh
INNER JOIN enrolments e ON sh.item_id = e.enrolment_id
LEFT JOIN statuses_history sh2 ON sh2.item_id = e.enrollment_id AND sh2.status_id = 50
WHERE sh.date_added > '2015-08-01 00:00:00' AND sh.date_added < '2015-08-20 23:59:59' AND sh.status_id = 57 AND (e.course_id = 2 OR e.course_id =7 OR e.course_id = 8 OR e.course_id = 9) 
GROUP BY e.campus_id, e.course_id;