"Tom Lane" <[EMAIL PROTECTED]> writes: > No, my proposed patch doesn't change that. It might be that we should > provide an "integer division" operator for NUMERIC, so that you can get > at the exact result of trunc(x/y).
I was also thinking that if the denominator had only factors of 2 and 5 we could calculate the precision to be precisely enough to maintain the original precision. Ie, nnnn/1000 should just give you n.nnn not n.nnn0000 and more importantly it should never round. -- Gregory Stark EnterpriseDB http://www.enterprisedb.com ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match