Question

SELECT 
  buttonID,
  live,
  title,
  image,
  (SELECT COUNT(*) 
    FROM otherdb.buttonclicks t2 
    WHERE t2.buttonid = t1.buttonID AND t2.`type`=0) as count 
FROM buttons t1
GROUP BY buttonID
ORDER BY live DESC, buttonID ASC;

We have 78 ad buttons on our site and have separated the stats from the main db for size reasons. We record every time one of the buttons is clicked on and I am working on a stats screen that can pull the information quickly.

The above query works but takes 4000 seconds (literally) to run, how can I get the same results much more efficiently.

Running the query as:

SELECT
  buttons.buttonID,
  buttons.live,
  buttons.title, 
  buttons.image, 
  count(buttonclicks.id) as count
FROM buttons INNER JOIN otherdb.buttonclicks ON buttons.buttonID = buttonclicks.buttonid
WHERE buttonclicks.type=0
GROUP BY buttons.buttonID
ORDER BY buttons.live DESC, buttons.buttonID ASC

is quicker but only picks up those buttons that have been clicked in the past three months (we archive the older clicks) because of the WHERE clause.

Solutions?

Answer 1

try this query.

SELECT tbl1.buttonID,
       tbl1.live,
       tbl1.title, 
       tbl1.image, 
       COUNT(tbl2.buttonid) as count
  FROM buttons AS tbl1
 INNER JOIN (SELECT buttonid FROM otherdb.buttonclicks b WHERE b.type=0) tbl2
    ON tbl1.buttonID = tbl2.buttonid
 GROUP BY tbl1.buttonID
 ORDER BY tbl1.live DESC, tbl1.buttonID ASC

I also observed that you have an image field. I think the query slow because of this since it’s a blob type that handles an image object.