Question

I have got a problem with a MySQL SQL-command.
I have got a table with 3 columns:

 id (int) | ip(string) | time(mysql timestamp)

I want to have a count with all rows grouped by time. I use the following code for it:

 SELECT DATE( TIME ) as date, COUNT( * ) as count
 FROM table
 WHERE TIME > ( NOW( ) - INTERVAL 10 DAY ) 
 GROUP BY DATE(  `time` )

Well, but now, I want to ignore double IPs. For example:
If i have put the following datas in the mysql table:

 ID | time                  | IP
 --------------------------------------
 1  | 2013-01-21 20:48:01   | 127.0.0.1  
 2  | 2013-01-22 20:48:01   | 127.0.0.1  
 3  | 2013-01-22 20:48:01   | 127.0.0.1
 4  | 2013-01-22 20:48:01   | 127.0.0.3  

I want only see the following datas:

 date       | count
 ------------------
 2013-01-21 | 1  
 2013-01-22 | 2  

What is the best way to do it? I’m not going to create in my application a loop which remove all duplicates because the table grows quickly and I will cost much resources if I do it in my application.

Answer 1

Try this:

SELECT DATE(`time`) as date, COUNT( DISTINCT IP ) as count
 FROM table
 WHERE `time` > ( NOW( ) - INTERVAL 10 DAY ) 
 GROUP BY DATE(  `time` )

SQL FIDDLE DEMO