topic Re: Exclude records with a '-' in the variable value in SAS Programming

Exclude records with a '-' in the variable value

SASMom2 — Sun, 02 Feb 2025 21:22:37 GMT

I need help to exclude records with a variable that has values other than numbers and characters. Below is an example.

I have a variable 'test' which has values like below.

Test
00*915174
00*915174
(00)-1234
00(9151.74)
00_915_174
00-9151-74
001234

In the above column, I would like to exclude all the values with special character such as (, ), -, ., * and space. I want to keep values such as '001234' or 'ABC123'.

Thank you

Re: Exclude records with a '-' in the variable value

PaigeMiller — Sun, 02 Feb 2025 21:29:13 GMT

It sounds to me like you could re-phrase the requirements to keep strings which contain digits or numbers, but not anything else.

And so this is a job for the FINDC function.

if findc(test,,'kad') then delete;

Re: Exclude records with a '-' in the variable value

SASMom2 — Sun, 02 Feb 2025 21:46:28 GMT

I feel so dumb asking this but still asking...what is 'kad' in your statement.

Thanks

Re: Exclude records with a '-' in the variable value

SASKiwi — Sun, 02 Feb 2025 23:01:07 GMT

@SASMom2 - Click on the FINDC documentation link provided by @PaigeMiller and you will find what 'kad' means.

Re: Exclude records with a '-' in the variable value

SASMom2 — Mon, 03 Feb 2025 00:41:37 GMT

Thanks you @SASKiwi

I figured it out.

Thanks you @PaigeMiller for your suggestion.

Re: Exclude records with a '-' in the variable value

Ksharp — Mon, 03 Feb 2025 00:57:05 GMT

data have;
input test $20.;
cards;
00*915174
00*915174
(00)-1234
00(9151.74)
00_915_174
00-9151-74
001234
ABC123
;
data want;
 set have;
 if notalnum(strip(test)) then delete;
run;

Re: Exclude records with a '-' in the variable value

webart999ARM — Mon, 03 Feb 2025 12:24:16 GMT

data want;
    set have;
    retain re_pattern;
    if _N_ = 1 then re_pattern = prxparse('/^[A-Za-z0-9]+$/');
    if prxmatch(re_pattern, strip(test)) then output;
run;

data want;
    set have;
    if not prxmatch('/^[a-zA-Z0-9]+$/', strip(test)) then delete;
run;