topic Strip Function Results in ODS and Base Reporting

Strip Function Results

bzh — Mon, 21 Dec 2015 20:27:08 GMT

In the SAS 9.4 documentation for Strip, under Details, I noticed this sentence:

If the value that is trimmed is shorter than the length of the receiving variable, SAS pads the value with new trailing blanks.

Can someone show me this in action? I would like to avoid this since it seems to go against what the Strip function is for.

Thanks!

Re: Strip Function Results

Astounding — Mon, 21 Dec 2015 20:32:12 GMT

A simple example:

if gender='F' then desc = 'Female';

else desc='Male';

desc = strip(desc);

The first statement defines DESC as being 6 characters long. If the last statement encounters "Male", that won't change the length of DESC. It will contain 6 characters: "Male" plus two trailing blanks.

Re: Strip Function Results

Reeza — Mon, 21 Dec 2015 20:33:14 GMT

It depends on what you're trying to do...certain things exclude trailing blanks by default, the length function is a good example. Note the differences between the length and lengthc variable.

data have;
length var1 $12. var2 $8.;

var2='Cat';
var2_lengthc=lengthc(var2);
var2_length=length(var2);
var1=strip(var2);
var1_lengthc=lengthc(var1);
var1_length=length(var1);
run;

proc print data=have;run;

Results:

                              var2_      var2_     var1_      var1_
                                 Obs    var1    var2    lengthc    length    lengthc    length

                                  1     Cat     Cat        8          3         12         3

Re: Strip Function Results

bzh — Mon, 21 Dec 2015 21:04:56 GMT

Astounding - in your example, if I were to make a Where clause for Male, would it be

where name='Male'

where name='Male '

I may be confused on whether or not these trailing blanks just take up storage or if they affect functionality. If one has a varchar type column with strings of varying length, it sounds like the shorter values will always have trailing blanks in a SAS data set because the Length must be set on the longest value (to avoid truncation).

Re: Strip Function Results

Astounding — Mon, 21 Dec 2015 21:07:18 GMT

You could use either. The software recognizes when it is comparing strings of unequal length. Without changing any of the values of your variables, it will automatically pad the shorter string with blanks as needed in order to make the comparison. Conveniently, you don't have to add the blanks but you will get the same results.