Question

I have a column contains urls. Some (but not all) urls share the same domain. Ex: aaa.yahoo.com, bbb.yahoo.com, ccc.yahoo.com. These urls are stored in a unique field. Also, the same table has an auto-increment primary key.

What I need to do and need help on it is: Query all the urls that share the same domain name. Note that, there is no exact domain name that I can put in the query, a lot of the shared domain names available in the database and I want to query them all.

EDIT:
This is the command I’ve tried but totally not sure as I don’t have that much experience in SQL:

SELECT domainname FROM DB.TABLE
WHERE SUBSTRING(domainname, LOCATE('.', domainname))
IN(
    SELECT   domainname
    FROM     DB.TABLE
    GROUP BY SUBSTRING(domainname, LOCATE('.', doaminname)) 
);

Answer 1

If I’m not understanding you incorrectly…

SELECT
    COUNT(`domainname`) AS 'occ'
    ,SUBSTRING(`domainname`,LOCATE('.',`domainname`)+1) AS 'domain'
FROM `the_table`
GROUP BY SUBSTRING(`domainname`,LOCATE('.',`domainname`)+1)
HAVING `occ`>1
ORDER BY `occ` DESC,`domain` ASC

The HAVING... line will filter out the domains that have only 1 occurence in the table.

---

EDIT:

Speaking on behalf of your comments… Because you can not know which is domain in the following examples (unless you test the values against a huge list of country codes, domain extensions etc)

• domain.co.uk
• subdomain.domain.co

you can not programmatically detect those. If you’re planning to rely on extension or the country extension, that won’t work either. What you can do is get a smart sorting and pick the ones by your eyes, that is unless you go about the above mentioned solution.

So, for example,

SELECT
    `id`
    ,`d` AS 'domain'
    ,IF(CHAR_LENGTH(`d`)-CHAR_LENGTH(REPLACE(`d`,'.',''))=2,
        SUBSTRING_INDEX(`d`,'.',1),'') AS 'f1'
    ,SUBSTRING_INDEX(SUBSTRING_INDEX(`d`,'.',-2),'.',1) AS 'f2'
    ,SUBSTRING_INDEX(`d`,'.',-1) AS 'f3'
FROM (
    SELECT
        `id`
        ,IF(CHAR_LENGTH(`domainname`)-CHAR_LENGTH(REPLACE(`domainname`,'.',''))>2,
            SUBSTRING_INDEX(`domainname`,'.',-3),`domainname`) AS 'd'
    FROM `the_table`
    GROUP BY `d`
) t
ORDER BY `f3`,`f2`,`f1`

might help you to see it. Note that you can not group the result by f2 and f3 because you might have

• example1.yahoo.com
• example2.yahoo.com

but you might also have

• yahoo.com.br
• google.com.br