topic Strange behaviour of %sysevalf with negative values of the macro variable in SAS Programming

Strange behaviour of %sysevalf with negative values of the macro variable

DrBoogie — Wed, 30 May 2018 08:00:08 GMT

Hi all, I have a question with respect to the %sysevalf. I use in a macro programm the %if-condition of the following form:

%if %sysevalf(&est)>0.4 %then %do;
....
%end;

It works fine if &est has positive values, for example 0.312 or 1.214.But if &est is negative, for example -0.091, then I become the error:

A character operand was found in the %EVAL function or %IF condition where a numeric operand is required.

I dont' understand why it could be important if the number in the macro variable is positive or not.

Has somebody seen this effect and has an explanation?

Thanks in advance.

Re: Strange behaviour of %sysevalf with negative values of the macro variable

RW9 — Wed, 30 May 2018 09:11:33 GMT

Should that not read:

%if %sysevalf(&est. > 0.4, boolean) %then %do;
....
%end;

Anyways, I highly advise avoiding doing data processing in macro - macro is after all just a text find/replace system and has virtually no concept of data.

Just o add, I think that as sysevalf does floating point calculations, when the var is de-referenced you see:

%syevalf(-0.091)

Which may look like like a calculation to the function, but incomplete. Just guessing here though, based on the manual:

Syntax

%SYSEVALF(expression<, conversion-type>)

expression: is an arithmetic or logical expression to evaluate.

Re: Strange behaviour of %sysevalf with negative values of the macro variable

DrBoogie — Wed, 30 May 2018 12:10:34 GMT

Thank You, with %sysevalf(&est>0.4,boolean) it works. However I don't really understand, why there is such a difference between positive and negative values. But SAS is sometimes quite hard to unterstand, so it's OK. 🙂

Re: Strange behaviour of %sysevalf with negative values of the macro variable

Astounding — Wed, 30 May 2018 14:47:34 GMT

In your original code, %SYSEVALF computed the value before the >

Then ....

Macro language needed to compare the results. It will automatically apply %EVAL when making the comparison in the %IF statement (comparing the results of %SYSEVALF to 0.4). However, %EVAL does not understand decimal points. So macro language automatically makes a character comparison, comparing the first character of the %SYSEVALF result to the "0" within "0.4".