Delete rows using conditions

[Marwa_Se]

Hello Everybody!

I have a dataset defined as bellow:

BANK_AGENCY           ID_Customer                Date                           Overtaking

          A                              0001                         10/02/2021                        3

          A                              0002                         12/03/2021                        3

          A                              0003                         13/04/2021                        3

          A                              0004                          25/05/2021                       3

          B                              0005                         18/01/2021                         2

          B                              0006                          25/02/2021                        2

          B                              0007                          03/06/2021                        2

Overtaking is the number of rows  to delete by agency and  ordering by date

For simplicity, I would like to get this result:

BANK_AGENCY           ID_Customer                Date                           Overtaking

          A                              0002                         12/03/2021                        3

          A                              0003                         13/04/2021                        3

          A                              0004                          25/05/2021                       3

          B                              0005                         18/01/2021                         2

Can you help me please ? Any suggestions are welcome 🙂

Thank u in advance

[Marwa_Se]

Sorry the overtaking is the rows number to keep and not to delete

[Kurt_Bremser · Accepted solution]

data have;
input BANK_AGENCY $ ID_Customer $ Date :ddmmyy10. Overtaking;
format date ddmmyy10.;
datalines;
A 0001 10/02/2021 3
A 0002 12/03/2021 3
A 0003 13/04/2021 3
A 0004 25/05/2021 3
B 0005 18/01/2021 2
B 0006 25/02/2021 2
B 0007 03/06/2021 2
;

proc sort data=have;
by bank_agency date;
run;

data want;
count = 1;
do until (last.bank_agency);
  set have;
  by bank_agency;
  if count le overtaking then output;
  count + 1;
end;
drop count;
run;

I take it that your expected result misses an observation for bank_agency = "B".

[Marwa_Se]

It works perfectly! thank you so much 🙂