Question

I have a mySQL query which takes a long time to process. I am querying a large table of IP ranges which relate to country codes to discover the country of origin for each IP in the url_click table. (IP database from from hxxp://ip-to-country.webhosting.info/)

It works brilliantly, albeit slowly.

Is there a more efficient way to write this query?

Table and output JPG: http://tiny.cx/a4e00d

SELECT ip_addr AS IP, geo_ip.ctry, count(ip_addr) as count
FROM `admin_adfly`.`url_click`,admin_adfly.geo_ip
WHERE INET_ATON (ip_addr) 
BETWEEN geo_ip.ipfrom AND geo_ip.ipto 
AND url_id = 165 
GROUP BY ip_addr;

Answer 1

The use of a function in the join between the two tables is going to be slower than a normal join, so you probably want to defer that particular operation as long as possible. So, I’d summarize the data and then join it:

SELECT S.IP_Addr, G.Ctry AS Country, S.Count
  FROM (SELECT ip_addr, COUNT(ip_addr) AS Count
          FROM admin_adfly.url_click 
         WHERE url_id = 165 
         GROUP BY ip_addr) AS S
  JOIN admin_adfly.geo_ip AS G
    ON INET_ATON (ip_addr) BETWEEN geo_ip.ipfrom AND geo_ip.ipto;

If you can redesign the schema and are going to be doing a lot of this analysis, rework one of the two tables so that the join condition doesn’t need to use INET_ATON().

Presumably, you have an index on the url_id column; that is the only one that will give you much benefit here.