BookmarkSubscribeRSS Feed
🔒 This topic is solved and locked. Need further help from the community? Please sign in and ask a new question.
Solph
Pyrite | Level 9

Hi,

 

I'm using %do year= 2002  %to 2014 to select cases and output resutls. When I select cases, I'm using two year values, say var1=2002 and  var2=2003, or var1=2003 and var2=2004, etc.,, What makes this complicate is the year 2 variable (with next year's value) is embended in a date format (e.g. var1='1apr2002'd and var2="31mar2003"d.)

 

I thought the following code would work, but it didn't (log is attached at the end). I'm using more years of data but here I have show 3 years.

- Am I using Symput wrong - is it that I can't put a macro variabe (&year) in the second argument?

- A related question. It worked if I simply just repeated the same line 13 times using different year values (such as lines 3-5 of the sample SAS code below). What didn't work is I wrote a series of output data set names  as DATA d2002-d2014 - SAS didn't regconize all datasets (d2002 d2003 d2004 d2005 etc.) and thus would not output data to some of the dataset. Is there a way to get around?

 

Sorry that I don't have a sample data for you to run. I'll just take your suggestions and apply to my data and report back.

 

Many thanks.

 

 

 

DATA d2002 d2003 d2004 subdata2002 subdata2003 subdata2004;
    SET mydata

    if effective_date <= '31mar2003'd and (end_date >= '1apr2002'd or end_date = .) then output d2002;  
    if effective_date <= '31mar2004'd and (end_date >= '1apr2003'd or end_date = .) then output d2003;  

    if effective_date <= '31mar2005'd and (end_date >= '1apr2004'd or end_date = .) then output d2004;  

   %macro get();
      %do year=2002 %to 2004;

          call symput('nextyr',1+&year.);

          if effective_date <= "31mar&nextyr."d and (end_date >= "1apr&year."d or end_date = .) then output subdata&year;
      %end;
   %mend;
    %get;
RUN;

 

 

LOG:

34 %macro get();
35
36 %do year=2002 %to 2003;
37 call symput('nextyr',1+&year.);
38 if effective_date <= "31mar&nextyr."d and (end_date >= "1apr&year."d or end_date = .) then output postal&year;
39 %end;
40 %mend;
41 %get;
NOTE: Line generated by the invoked macro "GET".
41 call symput('nextyr',1+&year.); if effective_date <= "31mar&nextyr."d and (end_date >= "1apr&year."d or end_date = .)
________________
77
41 ! then output postal&year;
MPRINT(GET): call symput('nextyr',1+2002);
WARNING: Apparent symbolic reference NEXTYR not resolved.
ERROR: Invalid date/time/datetime constant "31mar&nextyr."d.
MPRINT(GET): if effective_date <= "31mar&nextyr."d and (end_date >= "1apr2002"d or end_date = .) then output postal2002;
MPRINT(GET): call symput('nextyr',1+2003);
WARNING: Apparent symbolic reference NEXTYR not resolved.
MPRINT(GET): if effective_date <= "31mar&nextyr."d and (end_date >= "1apr2003"d or end_date = .) then output postal2003;
ERROR 77-185: Invalid number conversion on "31mar&nextyr."d.

1 ACCEPTED SOLUTION

Accepted Solutions
Tom
Super User Tom
Super User

You seem to be confusing CALL SYMPUT which is a data step statement that can only work WHILE THE DATA STEP IS RUNNING with %LET which can be used to assign macro variable values WHILE THE MACRO IS RUNNING.

 

%macro get();
%do year=2002 %to 2004;
  %let nextyr = %eval(&year +1);
          
if effective_date <= "31mar&nextyr."d and (end_date >= "01apr&year."d or end_date = .) then output subdata&year;      

%end;
%mend;


DATA d2002 d2003 d2004 subdata2002 subdata2003 subdata2004;
    SET mydata
    if effective_date <= '31mar2003'd and (end_date >= '01apr2002'd or end_date = .) then output d2002;  
    if effective_date <= '31mar2004'd and (end_date >= '01apr2003'd or end_date = .) then output d2003;  
    if effective_date <= '31mar2005'd and (end_date >= '01apr2004'd or end_date = .) then output d2004;  
    %get;
RUN;

View solution in original post

6 REPLIES 6
Tom
Super User Tom
Super User

You seem to be confusing CALL SYMPUT which is a data step statement that can only work WHILE THE DATA STEP IS RUNNING with %LET which can be used to assign macro variable values WHILE THE MACRO IS RUNNING.

 

%macro get();
%do year=2002 %to 2004;
  %let nextyr = %eval(&year +1);
          
if effective_date <= "31mar&nextyr."d and (end_date >= "01apr&year."d or end_date = .) then output subdata&year;      

%end;
%mend;


DATA d2002 d2003 d2004 subdata2002 subdata2003 subdata2004;
    SET mydata
    if effective_date <= '31mar2003'd and (end_date >= '01apr2002'd or end_date = .) then output d2002;  
    if effective_date <= '31mar2004'd and (end_date >= '01apr2003'd or end_date = .) then output d2003;  
    if effective_date <= '31mar2005'd and (end_date >= '01apr2004'd or end_date = .) then output d2004;  
    %get;
RUN;
Solph
Pyrite | Level 9

Got it and I just tested and it worked great. Thanks so much.

 

For the second question, If I write DATA sub2002 - sub2014, SAS will give error messages? Or I have to spell each of the output dataset names DADA sub2002 subs2003 sub2004 sub2005....? Is there a way not to list each one of them? 

 

DATA sub2002 - sub2014; 

   SET mydata

   %macro get();
   %do year=2002 %to 2014;
   %let nextyr=%eval(&year.+1);
   if effective_date <= "31mar&nextyr."d and (end_date >= "1apr&year."d or end_date = .) then output sub&year;
   %end;
   %mend;
   %get;
RUN;

 

Tom
Super User Tom
Super User

You cannot use ranges of datasets in a DATA statement.  You could use macro code to generate the names.

It is probably easier if you generate the whole data step with the macro.

%macro get(start,stop,inputds);
DATA 
%do year=&start %to &stop;
   sub&year 
%end;
;
   SET &inputds;
%do year=&start %to &stop;
   if effective_date <= "31mar%eval(&year+1)"d
     and (end_date >= "1apr&year."d or end_date = .) then 
     output sub&year
   ;
%end;

RUN;
%mend get;

%get(start=2002,stop=2014,inputds=mydata);
Solph
Pyrite | Level 9

Thanks. Saved the suggestions and will apply.

 

BTW it's such a great community to get help and response so quickly!

Reeza
Super User
This is essentially this problem:
http://www.sascommunity.org/wiki/Split_Data_into_Subsets

With the general consensus being - Don't do it 🙂
Solph
Pyrite | Level 9

Thanks. I understood the reasoning and generally I won't do it either. It's a special case that calls for it. 

SAS Innovate 2025: Save the Date

 SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!

Save the date!

How to connect to databases in SAS Viya

Need to connect to databases in SAS Viya? SAS’ David Ghan shows you two methods – via SAS/ACCESS LIBNAME and SAS Data Connector SASLIBS – in this video.

Find more tutorials on the SAS Users YouTube channel.

Discussion stats
  • 6 replies
  • 2159 views
  • 1 like
  • 3 in conversation