topic MACRO CODE in Statistical Procedures

MACRO CODE

robertrao — Fri, 28 Sep 2012 19:14:21 GMT

Hi Team, Could you please check if my macro needs to be made tidy or is this logic OK?

I have to subset observations into seperate cohorts(datasets) based on the codes

Sometimes using only 1 or 2 codes and sometimes using as many as 12 codes.

I did a macro because the A to J variables are common for all the "want" datasets and are coming from the same dataset "have"

I dont want to repeat the same steps 10 times

Would you recommend me to go ahead or u have any suggestions??

Thanks

%macro cohorts(want=,code1=,code2=,code3=,code4=,code5=,code6=,code7=,code8=,

code9=,code10=,code11=,code12=);

data &want;

set have(keep=A B C D E F G H I J );

where Proc_code in("&code1","&code2","&code3","&code4","&code5","&code6","&code7","&code8","&code9",

"&code10","&code11","&code12");

A1=put(A,agefmt.);

B1=compress(B);

B2=put(B1,$sexfmt.);

C1=put(C,BMIFMT.);

D1=D/3600;

D2=put(D1,OPTIMEFMT.);

E1=put(E,$DIA.);

F1=STRIP(F);

F2=put(F1,$SMOK.);

if substr(Pat,5) in("A", "B" ,"C","D","E") then pat=substr(Pat,1,4);else pat=Pat;

if Pat eq " " then delete;

run;

%mend;

%cohorts(want=headache,code1=32,code2=34,code3=36,code4=38);

%cohorts(want=fever,code1=57,code2=58);

%cohorts(want=nausea,code1=29);

%cohorts(want=jaundice,code1=630,code2=632,code3=633,code4=634,code5=635,code6=636,

code7=638,code8=639,code9=647,code10=677,code11=678,code12=679);

Re: MACRO CODE

Linlin — Fri, 28 Sep 2012 19:21:46 GMT

is your variable "Proc_code" a numeric or character variable?

Re: MACRO CODE

robertrao — Fri, 28 Sep 2012 19:27:43 GMT

Thanks for the reply LINLIN. Its charecter variable

Re: MACRO CODE

Linlin — Fri, 28 Sep 2012 19:32:36 GMT

I made some changes to your code. Please let me know if it works or not.

%macro cohorts(want=,code=)

data &want;

set have(keep=A B C D E F G H I J );

where Proc_code in(&code) and pat ne ' ';

A1=put(A,agefmt.);

B1=compress(B);

B2=put(B1,$sexfmt.);

C1=put(C,BMIFMT.);

D1=D/3600;

D2=put(D1,OPTIMEFMT.);

E1=put(E,$DIA.);

F1=STRIP(F);

F2=put(F1,$SMOK.);

if substr(Pat,5,1) in("A", "B" ,"C","D","E") then pat=substr(Pat,1,4);

run;

%mend;

%cohorts(want=headache,code="32","34","36","38");

Re: MACRO CODE

robertrao — Fri, 28 Sep 2012 20:00:56 GMT

Hi LINLIN,

There is an error which says:

46 %cohorts(want=graft,code="15732","15734","15736","15738");

-------

180

ERROR: All positional parameters must precede keyword parameters.

ERROR 180-322: Statement is not valid or it is used out of proper order.

Re: MACRO CODE

Linlin — Fri, 28 Sep 2012 20:06:02 GMT

then try:

%cohorts(want=headache,code="32" "34" "36" "38");

Re: MACRO CODE

Linlin — Fri, 28 Sep 2012 20:14:16 GMT

It should work. You code is similar to the example below.

data have;
set sashelp.class;
proc_code=put(age,2.);
run;

%macro cohorts(want=,code=);
data &want;
set have;
where Proc_code in(&code) ;
%mend;

%cohorts(want=headache,code="11" "12" "13" "38")
%cohorts(want=happy,code="15" "16")

Re: MACRO CODE

robertrao — Fri, 28 Sep 2012 20:18:30 GMT

Linlin,

It works perfectly fine now. Whats the trick there?

Why shouldnt we seperate them by a comma????

PLease explain

Thanks

Re: MACRO CODE

ballardw — Fri, 28 Sep 2012 22:01:41 GMT

The comma is no longer required for use with the IN operator though much sample code will use commas as it was previously require. I'm not sure which exact version made the change.

Re: MACRO CODE

Linlin — Fri, 28 Sep 2012 22:09:10 GMT

Comma separates the macro variables. when we use "%cohorts(want=headache,code="11" ,"12" ,"13", "38")".

macro variable code only takes one value (code="11").