>>I have a table that looks something like this:
>>
>>

>>groupnum          itemnum          quantity
>>1                  10                  1
>>2                  10                  1
>>1                  17                  5
>>2                  17                  5
>>1                  23                  2
>>1                  26                  1
>>2                  26                  1
>>1                  27                  3
>>2                  27                  4
>>1                  30                  0
>>2                  30                  0
>>2                  33                  1
>>1                  36                  3
>>2                  36                  5
>>

>>
>>The table isn't really sorted, but we order it this way in queries. As you can see, most itemnums are found in groups 1 and 2, and usually have the same quantity. The user wants a query that returns those itemnums found in both groups, but with different quantities. In the above example, such a query would return:
>>
>>

>>1                  27                  3
>>2                  27                  4
>>1                  36                  3
>>2                  36                  5
>>

>>There are probably a few ways to do this, involving sequences of queries, but I wonder what would be the most efficient.
>Why would

>1                  17                  5
>2                  17                  5

not included ?
>Cetin

I'm not sure if I understand your reply. The purpose of the query is to select pairs where the itemnum is the same, but the quantity is different for groupnum 1 and 2.