Question

I have a mySQL table that includes the following three columns: IDENT, TIME, and USER. All are type varchar. For example:

IDENT = someuser::24::myisp.net::us
TIME = 2012-11-05 1:45:42
USER = someuser

What I’m trying to do is to count the number of distinct IDENTs in the past 5 days. I tried the following, but it fails (not all distinct IDENTs pulled):

SELECT DISTINCT(IDENT)
     , (  UNIX_TIMESTAMP(now()) 
        - UNIX_TIMESTAMP(STR_TO_DATE(TIME, '%Y-%m-%d %H:%i:%s'))
       ) / 86400 AS numdays
FROM `PS_LOGIN_LOG` 
WHERE `USER` = 'someuser' 
GROUP BY IDENT
HAVING `numdays` <= 5

Answer 1

It’s not entirely clear what you mean by “in the past 5 days“, but something along the following lines should do the trick:

SELECT COUNT(DISTINCT IDENT)
FROM   PS_LOGIN_LOG
WHERE  USER = 'someuser'
   AND TIME >= CURRENT_DATE - INTERVAL 5 DAY
   AND TIME <  CURRENT_DATE + INTERVAL 1 DAY