topic Re: How to ignore non numeric portion of a character variable and convert it to numeric? in SAS Programming

How to ignore non numeric portion of a character variable and convert it to numeric?

saurabhc — Thu, 03 Oct 2013 22:16:29 GMT

I have a variable which has values like >=500, >75 in addition to truly numeric values in integers.

I need to convert it to a numeric variable while ignoring the =, <= etc.

The SAS input function does not work in this case.

Thank you.

Here is a sample set

data what;

input $lab;

datalines;

<=500

<75

=<100

2000

300

400

;

run;

Re: How to ignore non numeric portion of a character variable and convert it to numeric?

hsharmas — Thu, 03 Oct 2013 22:34:36 GMT

Hi, you can use a compress function here in ways it fits best and if the variable is non numeric use the input function to convert it to numeric.

Example: kk = compress(lab,"<= "); /*Here it will compress <= and blanks in the variable*/

kk = compress(lab,"1234567890","k") /* This will keep only 1234567890 values in the variable and it you have a decimal you can include . in the list*/

Use of input function: jj = input(kk,best12.); /***Assigning informat to the variable******/

format jj best12.; /*****Output/Display format***************/

Hope that helps!

Regards,

Somi

Re: How to ignore non numeric portion of a character variable and convert it to numeric?

saurabhc — Thu, 03 Oct 2013 22:47:25 GMT

I tried substr function. It is a longer way of doing this job.

if substr(result, 1, 2) in("<=","=<",">=")then results=substr(result,3,10);

else if substr(result,1,1)="<" then results=substr(result, 2,10);

else results=result;

results1=input(results,best12.);

Your code totally works though.

Thanks!

Re: How to ignore non numeric portion of a character variable and convert it to numeric?

hsharmas — Thu, 03 Oct 2013 22:56:00 GMT

K is a modifier and is used for keep. Also, in your method you could do it when obs are less as you can look and check where is the position of <>=, and impossible when their are thousands of obs.

Use this link:

SAS(R) 9.2 Language Reference: Dictionary, Fourth Edition

Rate me please if I answered you query.

Regards,

Somi