How to Retain Date value Using SAS datastep

Question

I have a SAS dataset with STS and CHGDATE. I would like to create a New Date variable NEWCHGDATE while retaining the same Date until the same STS=03 continues.

 ID               CHGDATE          STS
 101              11/22/2022        03
 101              11/23/2022        03
 101              11/24/2022        03
 101              03/08/2023        04
 102              01/11/2022        03
 102              01/12/2022        05
 102              02/12/2023        03
 102              02/15/2023        04


WANT
ID               CHGDATE          STS     NEWCHGDATE
101              11/22/2022        03       11/22/2022
101              11/23/2022        03       11/22/2022
101              11/24/2022        03       11/22/2022
101              03/08/2023        04       03/08/2023
102              01/11/2022        03       01/11/2022
102              01/12/2022        05       01/12/2022
102              02/12/2023        03       02/12/2023
102              02/15/2023        04       02/15/2023

For e.g In the above, ID 101 has STS=03 from 11/22 until 11/24 so I would like to retain the first Date 11/22/2022 until it reached STS=04 on 03/08/2023.

Here is my Query which is not giving me the expected result

  data want;
    set have;
    by id;
  if first.id then do;
    lag_sts = '';
    newchgdate = chgdate;
  end;
 else lag_sts = sts;
  if sts not in ('03') then newchgdate = chgdate;
   else newchgdate = ifn(missing(lag_sts),., lag(newchgdate));
 run;

Answer 1

Use a RETAIN statement to specify which variables will not get reset to missing at the top of the DATA Step implicit loop.

Use the LAG function to examine the prior sts value. Be careful about using LAG inside a conditional clause (which is not happening here). In this example the LAG is occurring in the test expression.

Example:

The test expression is a little tricky if you are new to SAS and DATA Step.

data have;
attrib id length=8 CHG_date length=8 informat=mmddyy10. format=mmddyy10. sts length=$3;
input id CHG_date sts ;
datalines ;
 101 11/22/2022 03
 101 11/23/2022 03
 101 11/24/2022 03
 101 03/08/2023 04
 102 01/11/2022 03
 102 01/12/2022 05
 102 02/12/2023 03
 102 02/15/2023 04
;

data want ;
  set have ;
  by id ;

  attrib new_CHG_date length=8 format=mmddyy10. ;
  retain new_CHG_date ;

  if sts ne '03' or lag(sts) ne '03' or first.id then 
    new_CHG_date = CHG_date ;
run ;

Answer 2

That is just normal BY group processing. But you will need to use the NOTSORTED keyword on the BY statement since the values of STS are not sorted. You just need to RETAIN the new variable and set it when you start a new ID or start a new STS that has value of '03'.

data have;
  input ID $ CHGDATE :yymmdd. STS $;
  format chgdate yymmdd10.;
cards;
101 2022-11-22 03
101 2022-11-23 03
101 2022-11-24 03
101 2023-03-08 04
102 2022-01-11 03
102 2022-01-12 05
102 2023-02-12 03
102 2023-02-15 04
;

data want;
  set have;
  by id sts notsorted;
  if first.id or (first.sts and sts='03') then newchgdate=chgdate;
  retain newchgdate;
  format newchgdate yymmdd10.;
run;

Result

Obs    ID        CHGDATE    STS    newchgdate

 1     101    2022-11-22    03     2022-11-22
 2     101    2022-11-23    03     2022-11-22
 3     101    2022-11-24    03     2022-11-22
 4     101    2023-03-08    04     2022-11-22
 5     102    2022-01-11    03     2022-01-11
 6     102    2022-01-12    05     2022-01-11
 7     102    2023-02-12    03     2023-02-12
 8     102    2023-02-15    04     2023-02-12