SQL - Way to find duplicate fields within multiple rows? - sql

Question

I'm wondering if there's a way to return duplicates of parts of rows.
IDTable setup:
ID# | Customer | EventID#
1 | Steve | 123
2 | Steve | 123
3 | John | 987
4 | John | 924
Since Steve and 123 appear twice together, I want to treat that as a 'duplicate' even though they have two different ID#'s. And if there's a 'duplicate', ideally I'd like to only return columns: ID#, Customer & EventID#. So for the above IDTable example, only return:
1 | Steve | 123
2 | Steve | 123
By running the following, it counts each ID + Customer + EventID# separately and returns all Count values as 1 (I'm using SQL Server 2008):
SELECT ID#, Customer, EventID#, COUNT({fn CONCAT(Customer,EventID#)})
FROM IDTable
GROUP BY ID#, Customer, EventID#
HAVING COUNT({fn CONCAT(Customer,EventID#)}) > 1
If I take out the ID# from the Select, it'l work but then we won't know what the ID#'s are.
EDIT:
I'm joining in the select columns from other tables. I initially left those out for simplicity sake by when trying to apply solutions below I'm getting confused. Apologies! Here's what is more in line with what I'm using:
SELECT A.ID#, C.Customer, E.EventID#
FROM IDTable A
INNER JOIN CustomerTable C
ON C.AccountID = A.AccountID
INNER JOIN EventTable E
ON E.AccountType = C.AccountType
WHERE C.StatusID = 'Active'

Answer 1

Self-Join should do the trick:
SELECT A.ID#, A.Customer, A.EventID#
FROM Table A
INNER JOIN Table A2 ON A.Customer = A2.Customer
AND A.EventID# = A2.EventID#
AND A.ID# <> A2.ID#
Edit for your joins:
You can still use a self-join, just with derived tables like so:
SELECT A.ID#, A.Customer, A.EventID#
FROM (SELECT ID#, Customer, EventID#
FROM IDTable A
INNER JOIN CustomerTable C ON C.AccountID = A.AccountID
INNER JOIN EventTable E ON E.AccountType = C.AccountType
WHERE C.StatusID = 'Active') A
INNER JOIN (SELECT ID#, Customer, EventID#
FROM IDTable A
INNER JOIN CustomerTable C ON C.AccountID = A.AccountID
INNER JOIN EventTable E ON E.AccountType = C.AccountType
WHERE C.StatusID = 'Active') A2 ON A.Customer = A2.Customer
AND A.EventID# = A2.EventID#
AND A.ID# <> A2.ID#
And cleaner with #TEMP:
SELECT A.ID#, C.Customer, E.EventID#
INTO #TEMP
FROM IDTable A
INNER JOIN CustomerTable C
ON C.AccountID = A.AccountID
INNER JOIN EventTable E
ON E.AccountType = C.AccountType
WHERE C.StatusID = 'Active'
;
SELECT A.ID#, A.Customer, A.EventID#
FROM #TEMP A
INNER JOIN #TEMP A2 ON A.Customer = A2.Customer
AND A.EventID# = A2.EventID#
AND A.ID# <> A2.ID#

Answer 2

Most version of SQL support window functions. The easiest way to solve this is:
select id, customer, eventid#
from (select i.*, count(*) over (partition by customer, eventid#) as cnt
from idtable i
) i
where cnt > 1;

Answer 3

SELECT i.*
FROM
IDTable i
INNER JOIN
(SELECT Customer, EventID#
FROM IDTable
GROUP BY ID#, Customer, EventID#
HAVING COUNT(*) > 1) t
ON i.Customer = t.Customer
AND i.EventId# = t.EventId#
There may be other ways of doing this and if you tag your specific rdbms (sql-server, oracle, mysql etc.) I am sure you will get additional answers but here is one way to do it which is to use your query (without the ID column) to identify the duplicates and then match it back to your original table via a inner join.

Answer 4

From #Aaron_D's answer I'd not join subqueries, instead you can do:
SELECT A.ID#, C.Customer, E.EventID#
FROM IDTable A
INNER JOIN IDTable B ON A.ID# = B.ID#
AND A.AccountID <> B.AccountID
INNER JOIN CustomerTable C
ON C.AccountID = A.AccountID
INNER JOIN EventTable E
ON E.AccountType = C.AccountType
WHERE C.StatusID = 'Active'
Since both CustomerTable AND EventTable are in last instance derived from AccountID it will work fine and will be faster.