topic Re: How to populate a value for all obs having same by variables in a data step in SAS Programming

How to populate a value for all obs having same by variables in a data step

knveraraju91 — Fri, 02 Dec 2016 01:58:06 GMT

Dear,

In my data set the following variable present.

if there is any change (variable CHG) by id and avisitn, greater than 10, then i need to populate (chgcat1='Max increase >10') for all obs having same id and avisitn

id avisitn chg test
1 2 8 sbp
1 2 9 sbp
1 2 7 sbp

1 3 12 dbp

1 3 4 dbp
2 1 10 sbp
2 1 11 sbp
2 1 5 sbp
2 1 4 sbp

output needed

id avisitn chg chgcat1
1 2 8
1 2 9
1 2 7

1 3 12 Max increase >10

1 3 4 Max increase >10
2 1 10 Max increase >10
2 1 11 Max increase >10
2 1 5 Max increase >10
2 1 4 Max increase >10

The chgcat1 is blank for first 3 obs because none of the obs with id=1 and avisitn=2 have chg greater than 10.

I think i can do this in two data steps. is there any way I can in one step as i have to calculate 12 chgcat. Thanks

Re: How to populate a value for all obs having same by variables in a data step

mkeintz — Fri, 02 Dec 2016 02:35:47 GMT

Read each group twice.

In the first group read look for chg>10

In the second group "re-read" output the records:

data want;

do until (last.avistn);

set have;

by id avistn;

if chg>10 then chgcat='Max Increase > 10';

end;

do until (last.avistn);

set have;

by id avistn;

output;

end;

run;

Re: How to populate a value for all obs having same by variables in a data step

Dan_B — Fri, 02 Dec 2016 03:24:59 GMT

I'd use the retain statement and have the data sorted. Then you can process in a single datastep as follows:

proc sort data=work.data1;
  by id avisitn descending chg;
run;

data work.data2;
  set work.data1;
  by id avisitn descending chg;
  retain chgcat;
  length chgcat $20;
  if first.avisitn then chgcat = '';
  if chg>10 then chgcat='Max Increase > 10';
run;

Re: How to populate a value for all obs having same by variables in a data step

Ksharp — Fri, 02 Dec 2016 11:03:52 GMT

You mean MAX > 10 to MAX-MIN > 10 ?


data have;
input id                    avisitn                   chg ;
cards;
1                       2                          8                     sbp
1                        2                          9                    sbp
1                       2                          7                     sbp
1                        3                          22                   dbp
1                       3                           4                      dbp
2                      1                           10                   sbp
2                      1                            11                   sbp
2                     1                              5                     sbp
2                      1                             4                     sbp
;
run;
proc sql;
select *,case when range(chg) gt 10 then 'Max increase >10' else ' ' end as f length=20
 from have
  group by id,avisitn;
quit;