Mysql query to find all rows that have the same values as another row

Question

My database contains rows that generally look like:

PersonItem
__________
id
personId
itemId

╔════╦══════════╦════════╗
║ ID ║ PERSONID ║ ITEMID ║
╠════╬══════════╬════════╣
║  1 ║      123 ║    456 ║
║  2 ║      123 ║    456 ║
║  3 ║      123 ║    555 ║
║  4 ║      444 ║    456 ║
║  5 ║      123 ║    456 ║
║  6 ║      333 ║    555 ║
║  7 ║      444 ║    456 ║
╚════╩══════════╩════════╝

I need to find all the actual records where the PersonId and the ItemId column match some other record in the database for those two columns....

| 1  |   123    |   456
| 2  |   123    |   456
| 5  |   123    |   456

| 4  |   444    |   456
| 7  |   444    |   456

How can I go about getting these results?

Answer 1

You can do joins to get around with duplicate records.

SELECT  a.*
FROM    TableName a
        INNER JOIN
        (
            SELECT  PersonID, ItemID, COUNT(*) totalCount
            FROM    TableName
            GROUP   BY PersonID, ItemID
            HAVING  COUNT(*) > 1
        ) b ON  a.PersonID = b.PersonID AND
                a.ItemID = b.ItemID

• SQLFiddle Demo

OUTPUT

╔════╦══════════╦════════╗
║ ID ║ PERSONID ║ ITEMID ║
╠════╬══════════╬════════╣
║  1 ║      123 ║    456 ║
║  2 ║      123 ║    456 ║
║  5 ║      123 ║    456 ║
║  4 ║      444 ║    456 ║
║  7 ║      444 ║    456 ║
╚════╩══════════╩════════╝

Answer 2

Something like this should do the trick:

SELECT P1.*
FROM PersonItem P1
INNER JOIN PersonItem P2 ON P2.ID <> P1.ID
AND P2.PersonId = P1.PersonId
AND P2.ItemId =   P1.ItemId

Answer 3

You need to find examples where the pair of personid/itemid appear more than once. In MySQL you can do this with a where clause and subquery:

select t.*
from t
where exists (select 1
              from t t2
              group by personid, itemid
              having count(*) > 1 and
                     t2.personid = t.personid and t2.itemid = t.itemid
             )

The above is standard SQL. MySQL also supports the multi-column in statement. So this can be written as:

select t.*
from t
where (t.personid, t.itemid) in (select personid, itemid
                                 from t
                                 group by personid, itemid
                                 having count(*) > 1
                                )

And alternative that I like, based on Eugene's answer but more efficient, is:

SELECT t.personid, t.ItemId, GROUP_CONCAT(t.ID)
FROM t
GROUP BY t.personid, t.ItemId
HAVING COUNT(*) > 1;

It does away with any joins, if you don't mind getting the ids as a list rather than as separate rows.

Answer 4

SELECT GROUP_CONCAT(p1.ID), p1.personid, p1.ItemId
FROM PersonItem AS p1
INNER JOIN PersonItem AS p2  ON 
    p1.ID<>p2.ID
    AND p1.personid=p2.personid
    AND p1.ItemId=p2.ItemId
GROUP BY p1.personid, p1.ItemId