Question

I have following MySQL query:

(SELECT c.Channel as name, count(*) as total_episode
 FROM (
     SELECT a.aid, a.vid 
     FROM videoItem v INNER JOIN aid2vid a USING(vid) 
     GROUP BY a.aid
 ) a1 INNER JOIN channelListingItem c USING(aid) 
 GROUP BY c.Channel
) 
UNION 
(SELECT c1.Channel as name, 0 as total_episode
 FROM channelListingItem c1 LEFT JOIN (
     SELECT c.Channel FROM (
         SELECT a.aid, a.vid 
         FROM videoItem v INNER JOIN aid2vid a USING(vid) 
         GROUP BY a.aid
     ) a1 INNER JOIN channelListingItem c USING(aid) 
     GROUP BY c.Channel
 ) c2 USING(Channel) 
 WHERE c2.Channel is null 
 GROUP BY name
);

Basically, what this statement does is to get the correct count episode in each channel & assign zero for channels w/o vid in consequent table (videoItem).

Note that

SELECT a.aid, a.vid
FROM videoItem v
INNER JOIN aid2vid a USING(vid)
GROUP BY a.aid

is duplicated twice and from explain this MySQL statement I don’t see MySQL re-use the query result.

+----+--------------+------------+------+----------+---------+---------+----------+------+---------------------------------+
| id | select_type  | table      | type | pos_keys | key     | key_len | ref      | rows | Extra                           |
+----+--------------+------------+------+----------+---------+---------+----------+------+---------------------------------+
|  1 | PRIMARY      | <derived2> | ALL  | NULL     | NULL    | NULL    | NULL     |  313 | Using temporary; Using filesort |
|  1 | PRIMARY      | c          | ALL  | idx_vid  | NULL    | NULL    | NULL     |  616 | Using where; Using join buffer  |
|  2 | DERIVED      | a          | ALL  | vid      | NULL    | NULL    | NULL     | 1015 | Using temporary; Using filesort |
|  2 | DERIVED      | v          | ref  | idx_vid  | idx_vid | 32      | db.a.vid |   10 | Using index                     |
|  3 | UNION        | c1         | ALL  | NULL     | NULL    | NULL    | NULL     |  616 | Using temporary; Using filesort |
|  3 | UNION        | <derived4> | ALL  | NULL     | NULL    | NULL    | NULL     |   28 | Using where; Not exists         |
|  4 | DERIVED      | <derived5> | ALL  | NULL     | NULL    | NULL    | NULL     |  313 | Using temporary; Using filesort |
|  4 | DERIVED      | c          | ALL  | idx_vid  | NULL    | NULL    | NULL     |  616 | Using where; Using join buffer  |
|  5 | DERIVED      | a          | ALL  | vid      | NULL    | NULL    | NULL     | 1015 | Using temporary; Using filesort |
|  5 | DERIVED      | v          | ref  | idx_vid  | idx_vid | 32      | db.a.vid |   10 | Using index                     |
|NULL| UNION RESULT | <union1,3> | ALL  | NULL     | NULL    | NULL    | NULL     | NULL |                                 |
+----+--------------+------------+------+----------+---------+---------+----------+------+---------------------------------+
11 rows in set (0.02 sec)

How do I refactor this MySQL statement? Also is there good refactor tool for MySQL statement?

Thanks.

Answer 1

This one seemed to work for me:

select Channel as name,count(distinct a1.aid) as total_episode
from channelListingItem c
left join
(
 select a.aid, a.vid 
 from videoItem v INNER JOIN aid2vid a USING(vid) 
) a1 on a1.aid = c.aid
group by Channel;

From what I can see the following query that you use twice in an inline view:

SELECT a.aid, a.vid 
FROM videoItem v INNER JOIN aid2vid a USING(vid) 
GROUP BY a.aid

Is being used to get a distinct list of aid and vid values that exist in both videoItem and aid2vid. I have replaced the GROUP BY in the inline view with a COUNT(DISTINCT) to achieve the same thing since you are not using any aggregate functions in the inline view part of the query.

I think you do not need to split the query into two parts joined by a union i.e. part 1 to get episode counts > 0 and part 2 to get episode counts = 0. This can be achieved in one GROUP BY.

Hope this helps!