topic Re: using proc sql in SAS Programming

using proc sql

Aayushi_17 — Mon, 19 Nov 2018 06:49:29 GMT

treat	COMPLETE	SAFETY	RAND
2	1	1	1
1	1	1	1
1	0	1	1
2	1	1	1

how to define study population using this ?

where

RAND = 1

SAFETY=1

PP=1

COMPLETE = 1

COMPLETE = 0

the output i want is

Randomized	xx (xx.x%)	xx (xx.x%)
Safety	xx (xx.x%)	xx (xx.x%)
Per Protocol	xx (xx.x%)	xx (xx.x%)
Completed	xx (xx.x%)	xx (xx.x%)
Discontinued	xx (xx.x%)	xx (xx.x%)

Re: using proc sql

PeterClemmensen — Mon, 19 Nov 2018 06:51:32 GMT

I don't understand your question? And why is PP not a part of your data?

Re: using proc sql

Aayushi_17 — Mon, 19 Nov 2018 06:55:30 GMT

yes PP is not part if my data and the output should be 0 for that

Re: using proc sql

error_prone — Mon, 19 Nov 2018 06:55:51 GMT

And why proc SQL? Is this homework? If so, please post what you tried so far.

Re: using proc sql

Aayushi_17 — Mon, 19 Nov 2018 08:32:56 GMT

i have found out till this
proc sql noprint;
select count(distinct pt)
into :n1 from dispositon
where rand = 1;
select count(distinct pt)
into :n2
from dispositon
where complete = 0;
select count(distinct pt)
into :n3
from dispositon
where complete =1;
select count(distinct pt)
into :n4
from dispositon where safety = 1;
quit;

i want the seperate count when trt is 2 and 1
like
trt 1 trt2
randomized 2 2
perprotocol 0 0

Re: using proc sql

RW9 — Mon, 19 Nov 2018 08:40:30 GMT

What you have there is a clinical report. Your company will have specific standards, macros etc. so you should really look into that with a colleague. The code associated with created those types of outputs are proprietary so cannot be shared. You can of course do something like:

proc sql;
  /* Create empty table */
  create table want (co1l char(200),trt1 char(200),trt2 char(200));
  /* Get treatment populations */
  select count(distinct usubjid) into :t1 from have where treat=1;
  select count(distinct usubjid) into :t2 from have where treat=2;
  /* Insert counts */
  insert into want 
  set col1="Randomized",
      trt1=(select strip(put(count(distinct usubjid),best.))||" ("||put((count(distinct usubjid) / &t1.)*100,4.1) from have where rand=1 and treat=1),
      trt2=(select strip(put(count(distinct usubjid),best.))||" ("||put((count(distinct usubjid) / &t2.)*100,4.1) from have where rand=1 and treat=2)
  insert into want...
;
quit;

So you can do that one table in one sql like that. Mostly though that will be a long winded way of doing it. You are better off learning the specific SAS functions like freq, means, summary, and using them to get counts. Then process the output in a datastep.