I recommend just using PROC APPEND, but assuming you only have dataset B you can use that to generate a LENGTH statement. proc sql noprint ;   select catx(' ',name,cats(case when (type=2) then '$' else ' ' end,length))     into :varlist separated by ' '     from b   ; quit; data addi;   length &varlist ;   .... code to populate variables  .... ; run; It would be better to include VARNUM in the metadata so that you could create the variables in the proper order. (Note generating macro variable DUMMY to suppress SQL message when order variable not included in the SELECT list.) proc sql noprint ;   select catx(' ',name,cats(case when (type=2) then '$' else ' ' end,length))        , varnum     into :varlist separated by ' '        , dummy     from b     order by varnum   ; quit; ... View more

What do you mean by "partitioned data"? Where is the data stored (SAS or a database - and if database: Which one and which version and is the table partitioned and how?). I found this link which is quite interesting and possibly what you have in mind: "Piping Between Data Step and Proc Sort on SMP Machine", http://support.sas.com/rnd/scalability/tricks/connect.html#pipds Message was edited by: Patrick ... View more

SAS is not a IDE tool like Eclipse or Netbeans . I am afraid it is too hard for sas .But you can use menu EDIT->REPLACE to replace the variable name , also you can use data step to processs the sas code(sas code is a txt file). But it is also limited for refactoring . Ksharp ... View more

Use fileexist() to replace fexist(). Ksharp ... View more

Without testing I would have expected that an error gets thrown because of overlapping ranges. But no: Proc Format digests these overlapping values. On my SAS9.3 Win64 installation the label from the lower range is taken independently in what order the ranges are defined with Proc Format (so it's always 'homie'). Wouldn't have thought about writing the format to a data set like RLC did. It's quite nice as it actually gives you an idea of what Proc Format does to deal with such an ambiguity. ... View more

It seems that it is difference between EDBIC and ASCII . ... View more

not sure why I feel I have to add this , but ... since no-one has pointed it out If your macro invokes SELECT  ....... INTO :mac_var1 then it might not be as _local_ as you expected %macro do_it( variable, data= sashelp.class ) ; proc sql noprint ; select max(&variable) into :mac_&variable  from &data ; quit ; %put _user_ ; %mend  do_it ; %macro caller ; %let mac_age = 1 ; %do_it( age ) ; %do_it( height ) ; %mend  caller ; option mprint ; %caller %put _user_; macro variables mac_age and mac_height are not global but mac_age is not local to %do_it ... View more

Or use %put _user_;  Check out. OR %symlocal %symglobl . %macro a; %let class="course_title"; data _null_; set sashelp.cars; run; %put %symlocal(class) %symglobl(class) ; %mend a; %a Ksharp ... View more

Yes, the only restriction is that you include complete statements, not partial statements. You cannot do data _null_; set have; if x > 0 then %include 'myFile.sas"; run; That doesn't work. But you can do: data _null_; set have; if x>0 then do; %include "myFile.sas"; end; run; But you must remember that the inclusion will occur even if x <= 0. The do block will not be executed, but the code from myFile.sas will be included. Hope that helps. PG ... View more

Also note ... because of the decimal point, %EVAL is forced to make a character comparison.  (As was noted, %EVAL can perform arithmetic with integers only.)  This expression would also return 1 because of the decimal point: %eval( 9. > 80) Since %sysevalf can perform non-integer arithmetic, this expression returns 0: %sysevalf( 9. > 80) ... View more

There are just a handful of cases.  Here are the ones I can think of off the top of my head. In a DATA step, CALL SYMPUT and SYMGET both expect to see the name of a macro variable.  So the ampersand should be omitted (although since it's in a DATA step, quotes would distinguish between a DATA step variable vs. a text string).  For example, both of these create a macro variable &SALES: varname='sales'; call symput('sales', 'Friday'); call symput(varname, 'Friday'); Other than that, the only case I can think of is the %superq function.  It also expects the name of a macro variable.  For example: data _null_; call symput('measure', 'S&P 500 Index'); run; %let measure = superq(measure); title "&measure"; This would quote the &P, so that resolving &measure does not trigger a search for the macro variable &P. ... View more

That's absolutely correct.  To put it in words ... With a dot, the expression resolves as Ksharp indicated.  Within &var.6, the dot delimits the name of the macro variable.  It tells macro language that VAR is the name of the macro variable, and 6 is just text to be appended to &VAR. Without a dot, the expression would resolve to &var6.  How would macro language know to look for &VAR, rather than &VAR6?  It doesn't.  This would trigger a search for &VAR6.  Since there is no such macro variable, it would generate an error. In general, keep asking!  Speaking for myself only, I'm more than willing to help when I see that you are putting in the work and trying hard to learn. ... View more