sql server - Increment Row Number on Group for regular withdrawal amounts -

i need create sql server query groups "common" sequential data , assigns unique id group. in example below have withdrawals made policy on regular basis.

create table #temp2 (policy_id int, wdl_date int, amount decimal(9,2));  insert #temp2 values(1001, 19881028, 190.00); insert #temp2 values(1001, 19881129, 190.00); insert #temp2 values(1001, 19881229, 494.89); insert #temp2 values(1001, 19890130, 494.89); insert #temp2 values(1001, 19890227, 494.89); insert #temp2 values(1001, 19890330, 494.89); insert #temp2 values(1001, 19890530, 525.00); insert #temp2 values(1001, 19890629, 525.00); insert #temp2 values(1001, 19890728, 525.00); insert #temp2 values(1001, 19890830, 525.00); insert #temp2 values(1001, 19930723, 51.00); insert #temp2 values(1001, 19931213, 190.00); insert #temp2 values(1001, 19940311, 190.00); insert #temp2 values(1001, 19940613, 190.00);  insert #temp2 values(1002, 19881028, 50.00); insert #temp2 values(1002, 19881129, 50.00); insert #temp2 values(1002, 19881229, 60.00); insert #temp2 values(1002, 19890130, 100.00); insert #temp2 values(1002, 19890227, 100.00); 

i create output such below column id populated query. note amounts sorted in date order should determine values in field id.

policy_id   wdl_date    amount  id ---------   --------    ------  -- 1001        19881028    190.00  1    1001        19881129    190.00  1 1001        19881229    494.89  2 1001        19890130    494.89  2 1001        19890227    494.89  2 1001        19890330    494.89  2 1001        19890530    525.00  3 1001        19890629    525.00  3 1001        19890728    525.00  3    1001        19890830    525.00  3 1001        19930723    51.00   4 1001        19931213    190.00  5 1001        19940311    190.00  5 1001        19940613    190.00  5  1002        19881028    50.00   1 1002        19881129    50.00   1 1002        19881229    60.00   2 1002        19890130    100.00  3 1002        19890227    100.00  3 

i've tried using dense_rank() populate id problem encounter 2 separate grouped amounts 190.00 (top 2 , bottom 3 records policy_id = 1001) need kept separate , dense_rank() statement considers both groups of 190.00 1 group. need grouping restart @ each new policy. code doesn't produce required output.

select *,        dense_rank() on (order policy_id, amount) id     #temp2 order policy_id, wdl_date;     

can provide guidance on this?

to 190.00s need this...

select *,        dense_rank() on (partition policy_id order amount) id     #temp2 order policy_id, id;  

your desired output has discrepancies when order amount in ranking function value 51.00 should first value not 190.00