Question

I’m pulling a grouping of comment types and their totals from a mySQL database. The result set looks like:

ID – NAME – TOTAL
0 – GENERAL – 4
3 – MEANINGFUL – 9
4 – MISC – 5

In my output to the end user, however, I want to sum the totals for GENERAL and MISC, IDs 0 and 4. How should I efficiently do this? The problem I’m seeing is that ID 0 might not be in every result and the same thing for ID 4.

Any suggestions?

My query is:

SELECT a.fk_type_id as ID, count(a.fk_type_id) as TOTAL, b.vch_name as NAME
FROM comments a
LEFT JOIN comments_types b ON a.fk_type_id = b.id
WHERE a.entry = X
GROUP BY a.fk_type_id

So my end-user result should look like:

GENERAL/MISC – 9
MEANINGFUL – 9

Answer 1

I would make it easy on you and MySQL and split it out like this:

(SELECT a.fk_type_id as ID, COUNT(a.fk_type_id) as TOTAL, b.vch_name as NAME
FROM comments a
JOIN comments_types b ON a.fk_type_id = b.id
WHERE a.entry = X AND a.fk_type_id <> 0 AND a.fk_type_id <> 4
GROUP BY a.fk_type_id)
UNION ALL
(SELECT a.fk_type_id as ID, COUNT(a.fk_type_id) as TOTAL, 'GENERAL/MISC' as NAME
FROM comments a
JOIN comments_types b ON a.fk_type_id = b.id
WHERE a.entry = X AND (a.fk_type_id = 0 OR a.fk_type_id = 4)
GROUP BY a.fk_type_id)

I changed your LEFT JOIN to JOIN because we’re only counting comments with a type.

You might improve this query, assuming an index on fk_type_id, by making ‘MISC’ type ‘1’ instead of type ‘4’, so you could do something like a.fk_type_id < 2 instead of (a.fk_type_id = 0 OR a.fk_type_id = 4) and a.fk_type_id > 1 instead of a.fk_type_id <> 0 AND a.fk_type_id <> 4.

If you’re utilizing indexes instead of scanning rows, performing two separate queries that return different results won’t affect performance.