Thanks Tom's mprint, I might have got the reason: *%abc does not contain macro statements, so it was put in the word queue where the statement is regarded as SAS statement and * does work then to comment out %abc. However, %let is a standard macro statement so it got executed immediately. Please straighten me out if it still is not much correct. ... View more

Thanks for the help. I am trying to understand why * before %abc does block the execution of the macro, whereas it does not in  *%let x=%abc;  I am not sure if ";" plays any role after *%abc since it does not change the result at all. ... View more

Thanks Tom. Combining with the reply from ballardw, I finally got what is going on: The unmasked comma in &final_class is regarded as one parameter separator by IFC in the macro facility, then only the first segment is returned since the condition evaluation is positive.   ... View more

Agree. But still, how come the returned value is truncated? If I simply replace the macro variable with its true value, the truncation is not happening:   %put WARNING: %sysfunc(ifc(%length(abv,bdev),%str(,abv,bdev),)); WARNING: ,abv,bdev   Therefore, it seems %str(,&final_class) coming with something unexpected inside of the %sysfunc.  ... View more

Two more questions on this thread:     1) I am curious why TRANWRD function does not base the resultant value length on the target string, i.e. the variable ALL in the example.  Different with other string functions, TRANWRD must have one and only one target.  Is it very natural for the function to return a new value with the equal length, if its length is not yet defined?  With this assumption, the variable NEW was not defined originally in the DATA-step.      2) How can TRANWRD be used as operands of another function, such as COALESCEC function. I.e. in order to avoid the truncation issue, must the resultant value be given to a pre-defined variable first?  Here is the program where the truncation was originally found:     proc sql;         select coalescec(tranwrd(o.all,'ONE','TWOO'),t.all) as newcode length=500        from onedata as o natural left join                   twodata as t        ;     quit;   Clearly, truncation in TRANWRD still happens without being affected by the good length of NEWCODE.  It will work that a temporary variable temp_var is derived first based on tranwrd(o.all,'ONE','TWOO'), then coalescec(calculated temp_var, t.all).  However, the coding is a little clumsy. Is there another way to make it better? ... View more