Question

My analytics team uses a query like this:

SELECT 
    SUM(ul.upload_bandwidth) + SUM(dl.download_bandwidth) AS total 
FROM 
    upload_facts ul, download_facts dl
WHERE 
    ul.date BETWEEN '2011-09-01' AND '2011-09-30' AND 
    dl.date BETWEEN '2011-09-01' AND '2011-09-30';

This is taking a very long time, over 12000 seconds, due to the “join” that is occurring.

Doing the sum from either table separately takes just a few seconds. I am considering putting a sum per day into a separate table to speed this up. However, I believe that it should be easier than that.

I want to eliminate the join and use subqueries to do this, however I’m not 100% sure how. I tried the following, but it did not work.

SELECT 
    (select upload_bandwidth from upload_facts where date 
      BETWEEN '2011-09-01' AND '2011-09-30') 
    + 
    (select download_bandwidth from downloaded_facts where date 
      BETWEEN '2011-09-01' AND '2011-09-30');

Answer 1

SELECT
    (SELECT SUM(`upload_bandwidth`) FROM `upload_facts` WHERE `date` 
      BETWEEN '2011-09-01' AND '2011-09-30')
    +
    (SELECT SUM(`download_bandwidth`) FROM `download_facts` WHERE `date` 
      BETWEEN '2011-09-01' AND '2011-09-30')
AS `total`

That should do it.