How can I remove all characters from string starting at first non-alpha character?

Question

This would have been a lot easier if not for certain situations.

Sample data:

KENP989SD
KENP913E
KENPX189R
KENP913

What regular expression can I use to remove all characters from the string starting at the first non-alpha character? Basically, I want to find the first non-alpha character and chop everything off after that regardless of char type.

After regex is applied, these data should be:

KENP
KENP
KENPX
KENP

Answer 1

s/([A-Za-z]*).*/$1/

... will work. It's not necessarily the best way of doing it, but it's a general case replace.

It only works if you just want alpha characters

Answer 2

Maybe this:

s/(?<=^[A-Z]+).*//

Uses look-behind to replace everything after the starting alphas with blank.

Add an i flag for case-insensitive if necessary:

s/(?<=^[A-Z]+).*//i

Answer 3

$s =~ s/[^a-zA-Z].*$//;

Literally, find the first non-alpha char and chop everything off starting from it.

Answer 4

NOTE: I think Igor's is more efficient.

$str =~ s{^([A-Z]+).*}{$1};

Add the 'i' flag for case-insensitive matches

$str =~ s{^([A-Z]+).*}{$1}i;

Answer 5

You phrased the request 2 ways:

1. Get all the alpha chars off the front of these strings
2. Find the last alpha char and chop everything off after

While the result is the same given your sample strings, I've found it pays to be more careful with regexes. So, I'd take the first item above as the real requirement, and write it as:

$str =~ s/^([a-z]*)[^a-z].*/$1/i;

The advantage in my mind is that unexpected strings (like "7KENP989SD") should result in a null string after substitution, instead of something unexpected like "7KENP". Of course, maybe that is what you wanted...

Answer 6

Here's my go at it.

/^([A-Za-z]).$/

---

EDIT I like Igor's approach better than mine ..

---

code:

#!/usr/bin/perl
#
# http://stackoverflow.com/questions/507941/perl-regex-remove-all-characters-from-string-after-last-alpha-character
#
use strict;
use warnings;
for my $string (<DATA>){
    $string =~ /^([A-Za-z]*).*$/;
    print "$1\n";
}
__DATA__
KENP989SD
KENP913E
KENPX189R
KENP913

Answer 7

If you don't need to modify the input line itself, I use this a little more:

my ( $alpha_prefix ) = ( $input_line =~ /^(\p{IsAlpha}*)/ );

Most of my variables are lexicals in a vast majority of cases, so a few more don't hurt and keeps me from possibly misrepresenting input. Plus, it passes taint.

Answer 8

s/\P{Alpha}.*// works for me fine:

perl -pe 's/\P{Alpha}.*//' <<EOF
KENP989SD
KENP913E
KENPX189R
KENP913
EOF