How to use regex to test variable in awk?

Question

I want to test if a script argument is only composed of letters. here is the script :

 BEGIN {
        VALUE=ARGV[1];
        if (VALUE ~ /[A-Za-z]/) {
           print VALUE " : Ok only letters";
        }
        print VALUE;
}

it seems that it matches every string with at least one letter :

tchupy@rasp:~$ awk -f s.awk file 111
value = 111
tchupy@rasp:~$ awk -f s.awk file  @@@
value = @@@
tchupy@rasp:~$ awk -f s.awk file  aaa
aaa : Ok only letters
value = aaa
tchupy@rasp:~$ awk -f s.awk file  1a1
1a1 : Ok only letters
value = 1a1
tchupy@rasp:~$ awk -f s.awk file  a1a
a1a : Ok only letters
value = a1a
tchupy@rasp:~$ awk -f s.awk file  1@1
value = 1@1

I tried to use the match() function, but I've got a syntax error at or near [ when I try to use [A-Za-z] regex.

Thx

Answer 1

Your test will be true if the variable contains at least one character from your character class. To test if the variable only contains characters in your character class, you need to match from the beginning (^) to the end of the script:

BEGIN {
    VALUE=ARGV[1];
    if (VALUE ~ /^[A-Za-z]+$/) {
       print VALUE " : Ok only letters";
    }
    print VALUE;
}

Or more concisely:

BEGIN {
    print ARGV[1] ": " (ARGV[1] ~ /^[A-Za-z]+$/ ? "OK" : "BAD")
}