topic determining length count for observations within the same variable in SAS Procedures

determining length count for observations within the same variable

sophia_SAS — Thu, 15 Mar 2012 12:35:35 GMT

SAS Community-

Using a large dataset, there is the numeric variable - 'SubjectID'. Within 'SubjectID' the observations including a number of different lengths. For example some SubjectID observations have 6 numbers while others have 7 or 8 numbers.

SubjectID

234583

5546492

5546489

234577

68907734

68907736

234581

Please advise on a SAS function that will inform me on how many Subject IDs have 6 for a number length, 7th for a number length and 8 for a number length.

Eventually, I will use this information to for a substring.
Thanks!

determining length count for observations within the same variable

Haikuo — Thu, 15 Mar 2012 12:41:15 GMT

Hope this will get you started:

data have;

input SubjectID $8.;

len=length(subjectid);

cards;

234583

5546492

5546489

234577

68907734

68907736

234581

;

proc freq data=have;

table len;

run;

Regards,

Haikuo

determining length count for observations within the same variable

sophia_SAS — Thu, 15 Mar 2012 13:57:06 GMT

Haikuo-

One quick follow-up question, how do I do this step without using the input statement?

I want to use a dataset currently located in my work library. According to proc contents for this dataset, the subject id has the variable listed as numeric with 6 as the length (though I know there are subject ids with up to 9 numbers).

I'm sure this is a very simple SAS 101 question. I just can't seem to get my head around it.

Thanks.

determining length count for observations within the same variable

Haikuo — Thu, 15 Mar 2012 14:08:02 GMT

I hope the following will help you with your problem:

/*This to input the numberic value*/

data have;

input SubjectID ;

cards;

234583

5546492

5546489

234577

68907734

68907736

234581

;

data want;

set have;

len=length(strip(subjectid));

/*This to get the occupied length*/

run;

proc freq data=want;

table len;

run;

determining length count for observations within the same variable

ballardw — Thu, 15 Mar 2012 14:12:36 GMT

If the values are numeric the length refers to the number of bytes used to store the value. The range for length of numerics is 3 to 8. This might do what you need.

data len;

set <your dataset>;

len = length (put(SubjectId,f10.0));

run;

proc freq;

table len;

run;