You need to educate yourself using this program as a learning tool.
First, picture what you want the DATA step to look like (with no macro language at all).
Second, add this statement and then re-run:
options mprint;
You will be able to see what the DATA step actually looks like, and compare it to your picture of what you wanted it to look like.
... View more
You're mixing macro and data step logic and with out the rest of the code or any sort of sample data or a clear question it's going to be really hard to help you.
... View more
Also check this two statements:
define bnrind / display style={just=left cellwidth=2.1cm} "Baseline Reference Range Indicator";
define A2N /group order=data style={just=left cellwidth=2.1cm} "Baseline Reference Range Indicator";
you only need the A2N, as this us used in the COLUMN statement
... View more
I would ask how and why you create a blank macro variable in the first place.
Since your code snippet is an incomplete macro referencing macro variables without any example value, specifying which macro variable is blank when, or showing any code related to actually creating a data set (such as "data ¯odatasetname;" I can't make any reasonable targeted suggestions in response to your specific question.
... View more
Why would you want to do that? You would have to determine the array sizes from the number of distinct values for dose_fmt, and use quite a lot of macro programming and/or call executes to set the final datastep up. Proc transpose is much more flexible, as you have seen, and does most of that work on its own.
PS of course one can drive screws in with a hammer, but I much prefer to use a screwdriver. See Maxim 14.
... View more
Applying the format to primreas will not change the order of a proc sort. Procedures that can order by formatted values usually have options to suppress this.
Or you create a new variable and leave primreas itself untouched.
No merge is needed to apply a format.
If you still have problems, supply an example for the expected result, based on your example data.
... View more
This:
id paramcd seq;
Wont work. Update here:
data Id_adds (keep=studyid usubjid aval avalc paramcd seq);
set anl2;
by studyid usubjid paramcd aval; retain seq;
if first.paramcd then seq=1; else seq+1; paramcd=cats(paramcd,put(seq,best.));
run;
Then put:
proc transpose data=Id_adds out=Id_trans(drop=_name_ _label_) ;
by studyid usubjid;
id paramcd;
var aval;
run;
Ie make the variable paramcd cd unique by adding the incremental number to it (when you have the sequence correct).
... View more
You might want to try
column rpt_varn rpt_var actarm rpt_cat, count dummy ;
The , after rpt_cat says that you want to summarize count for each level of rpt_cat.
Since you didn't actually have any values associated with rep_cat it was, I believe, defaulting to an N statistic as the variable rpt_cat is character and that is the default statistic for character (hard to sum character variables...)
... View more
@vraj1
Looking at the data and code you've now provided again: May be below gives you what you're after.
proc transpose data=Id_adds out=Id_trans(drop=_:);
by subjid;
id paramcd seq;
var avalc;
run;
proc sql;
create table want as
select
idt.*,
l.aval
from
id_trans as idt
left join
(select subjid, aval, avalc from anl2) as l
on idt.subjid=l.subjid
order by l.aval, idt.subjid
;
quit;
... View more
/Sorry, the code you post:
dsreas_num_%scan(&ArmList.,&k., #)=%scan(&ReasLst., &i.);
Does not match anything else you posted before. What is armlist, reaslist &i etc. Look at what I have posted. Get your data into a usable way, do datasteps to get it as you want then merge the data together, forget macros and other things this is basic data manipulations.
... View more
@vraj1 wrote:
If it fulfills all conditions then it should be multiple. But what i get is multiple obs of subjids for almost most of them
Then they all fulfill multiple conditions. Maxim 3: Know your data.
... View more
It is easier to retrieve that from dictionary.tables:
proc sql;
create table test as select * from dictionary.tables where libname='SASHELP' and memname = 'CLASS';
create table test2 as select modate from dictionary.tables where libname='WORK' and memname = 'TEST';
quit;
... View more
@vraj1
Something like below should do.
data have;
infile datalines truncover;
input var $20.;
datalines;
8.3
22 *
6.7 H *
120 H *
34 L *
;
run;
/* option 1 */
data want1;
set have;
if prxmatch('/(h|l)\s*\*$/oi',strip(var));
run;
/* option 2 */
data want2(drop=_:);
set have;
_var=compress(var);
if upcase(substrn(_var,lengthn(_var)-1)) in ('L*','H*');
run;
/* option 3 */
data want3;
set have;
if strip(scan(var,-1,'lh','i')) ='*';
run;
Preferred option: Option 3
... View more
Out of interest, why? Neither of those two datasets you provide as example are of much use to anyone. Could just be the test data you provide, however there is no way from either of those to tell if 43535 should belong to time or jack for instance. I would suggest sorting out the data before getting to this stage, and assign proper ids and such like so that you can correctly identify a unique datapoint.
... View more
