Application of a scalar function across a varying number of records within an RDBMS using the RDBMS's SQL capabilities. This is accomplished by determining a last row in a group of rows, sorting the rows and passing the sorted rows to a scalar function, where the scalar function accepts each row and stores the accepted rows in a temporary file space. When the last row is encountered, the aggregate value is calculated using all of the rows in the group, and then a filter is applied on the rows so that only a row with the aggregate value is returned.