You have tab characters in your data but you haven't told SAS that they are there.  Try adding dlm='09'x to your infile statement. ... View more

SCAN can work with multiple delimiters, so you could for example amend RW9's code as follows:      do until (scan(text_string,i," ,;/")=""); /* multiple delimiters */           var{I}=scan(text_string,i," ,;/");           I=I+1;      end; and it would then split at every blank, coma, semicolon or forward slash. ... View more

Perhaps it was a homework question?  This is Google's cache of what mmm3902 originally posted: ---------------------------------------------------------------------------------------------------------------- Data1: Company_Code Event Event_Year Data2: Company_Code Year Statistic Data1 Company_Code Event Event_Year 001 1 2000 002 0 . 003 1 1998 004 0 . Data 2 Company_Code Year Statistic 001 1999 60 001 2000 55 001 2002 60 001 2003 36 001 2004 12 001 2005 12 003 1999 60 003 2000 60 003 2002 55 004 2000 36 004 2005 60 I need to program where for each Company_code in Dataset1; I look for the same Company_code in Dataset2.The following should all go into Dataset1: Compare Event_year (Dataset1) to Year (Dataset2). If Event_year (Dataset1) > Year (Dataset2); INDICATOR =1; else INDICATOR=0; If Event_year(Dataset1) > Year (Dataset2); INDICATOR_PREV =1; else INDICATOR_PREV=0; For INDICATOR =1, calculate SUM(Statistic) For INDICATOR_PREV, calculate SUM(Statistic) For 3 years prior to each event for each company in Dataset 1, calculate corresponding average of STATISTIC. For 3 years after each event in Dataset for each company in Dataset 1, calculate corresponding average of STATISTIC. Any help on this would be appreciated! I've taken on a bit more than I could handle right now and am hoping for some guidance. Thanks! ... View more

ISO 8601 formats are available in v9 and are documented here: SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition But personally I find them rather fiddly and I think Hima's suggestions are good. ... View more

Actually in the dataset created, the DEPT is a numeric variable so I compare it with number 10. That's a data step variable, not a macro variable; they are not the same thing. But even if you do have a macro variable called DEPT that has a numeric value, your code %if dept<10 still fails for the reason I gave above; you need to instruct the macro compiler to resolve the variable by coding %if &DEPT < 10 . ... View more

This statement: %if dept<10 instructs the macro compiler to compare the character string dept with the number 10.  This is why you got the message "A character operand was found ... where a numeric operand is required." Is there any need to use macro language at all? The data step can probably do everything you need. ... View more

Remember that SUBSTR can be called with a variable as the third term, in which case there is no way of knowing at compile time whether the field will be truncated at run time - it's going to be data dependent.  I think we can infer that the compiler has been written to handle that as the general case, and hasn't had a special override added to say 'but if the user has specified a literal in position 3, check the table definition because we might not need the warning message after all.' ... View more

Defining a view and using that to drive the update gets rid of the warnings... proc sql;   create view aview as   select a.*   , substrn(md5(name || put(height,z3.) || put(weight,z6.2) || sex),1,16) length=16 as hash_key  format=$hex32.   , substrn(catx('|', name, put(height,z3.), put(weight,z6.2), sex),1,32) length=32 as cats_var    from sashelp.class a   ;   insert into asql   select a.*    from aview a   ; quit; Hope this circumvention is helpful. ... View more