Remove same values rows over different column mysql - php

Question

I 'm stuck on a very basic problem, I want to skip the row which has duplicate values over three columns.
Table feeds
id,type,user_id,friend_id,date
1, 'friend', 4 , 5 , 5/5/2010
2, 'friend', 5 , 4 , 5/5/2010
Now this is how the data is saved (I can't change the saving module)
since both have same thing, so I want to pick them only as a 1 row not 2.
I don't want to validate and remove it at PHP end, b/c if I'll do at PHP the pagination would be disturb
Edit:
Table Structure
create table `feed` (
`id` double ,
`type` blob ,
`user_id` double ,
`type_id` double ,
`date` datetime
);
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('78','friends','1314','1313','2012-09-03 19:48:14');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('79','friends','1313','1314','2012-09-03 19:48:14');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('80','friends','1314','1312','2012-09-03 19:49:07');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('82','friends','1313','1312','2012-09-03 19:49:09');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('84','friends','1315','1312','2012-09-03 19:49:24');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('86','friends','1315','1313','2012-09-03 19:49:33');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('87','friends','1313','1315','2012-09-03 19:49:33');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('97','friends','1317','1312','2012-09-03 19:55:06');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('99','friends','1313','1317','2012-09-03 19:56:01');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('100','friends','1317','1313','2012-09-03 19:56:01');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('101','friends','1315','1317','2012-09-03 19:56:58');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('102','friends','1317','1315','2012-09-03 19:56:58');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('104','following','1313','1193','2012-09-03 19:59:39');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('105','following','1313','1308','2012-09-03 19:59:51');
insert into `feed` (`id`, `type`, `user_id`, `type_id`, `date`) values('106','following','1313','1098','2012-09-03 19:59:58');

Answer 1

And here is the ultimate solution! If the same friendship pair (reversed) exists it only takes the one where user_id>friend_id.
SELECT DISTINCT type, user_id, friend_id, date
FROM table t1
WHERE t1.user_id > t1.friend_id
OR NOT EXISTS (
SELECT * FROM table t2
WHERE t1.type=t2.type AND t1.date=t2.date
AND t2.user_id = t1.friend_id AND t2.friend_id = t1.user_id
)

Answer 2

The function you are looking for is DISTINCT(). That allows you to group the table by that column and remove duplicates.
Just make sure in your select statement that you also select the other field columns as well:
SELECT id, DISTINCT(type), user_id, friend_id, date FROM TABLENAME

Answer 3

You could perform a JOIN on the table itself, which should do what you are looking for.
Something like this might do the trick:
SELECT * FROM tableName a
JOIN tableName b ON a.user_id = b.friend_id

Answer 4

You should get rid of those duplicate records by running:
DELETE FROM table t1 WHERE
EXISTS (
SELECT * FROM table t2
WHERE t1.type=t2.type AND t1.date=t2.date
AND t2.user_id = t1.friend_id AND t2.friend_id = t1.user_id
AND t1.user_id > t2.user_id
)
Then you can get distinct friendship records with:
SELECT DISTINCT type, user_id, friend_id, date FROM table
Also, you can write a DB trigger that fires on INSERT, that checks if a duplicate (with reversed user_id and friend_id) already exists. If exists it does ROLLBACK, else allow INSERT.