>SELECT T1.fld1,T1.cfld1,COUNT(*) FROM test2 T1 JOIN test2 T2 ON T1.fld1=T2.fld1 AND T1.cfld1=>T2.cfld1 GROUP BY 1,2
>
>This is good if fld1,cfld1 is a primary record tag for the table.
>

I noticed that you are joining TEST2 with TEST2 (the same child table). The tables were very simple representation of parent-child relation and I also need other columns from the header (TEST1).

However, I'm keeping this query. I know it will come in handy. Thank you.