Individually, you can match control characters using the syntax "\cX" where X is A for ^A B for ^B and "\c@" for a null byte chr(0) etc.

You can also use "\0" or "\00" or "\000" (octal) or "\x00" (hex) all of which will match '^@'.

For your particular problem, you would probably be much better off using tr to remove the unwanted characters as in

$s = "\cA000014\cAB0000BVE1I\c@";

print $s;
?000014?B0000BVE1I

$s =~ tr[\x20-\x7f][]cd;

print $s;
000014B0000BVE1I
[download]

Breifly, that says "translate anything in the range hex 20 (dec:32 <space>) through hex 7f (dec:127, <rubout>)".

But then the switches /c say's to complement the list, so now x20-x7f will be left alone and anything else will be transformed. The /d switch says delete anything that matches the (complemented) list. (See the tr// operator in perlop under "Regexp quote-like operators").

Update: Using tr[\x20-\x7e][]cd; so that rubouts are also deleted might be more appropriate. Either range will delete the extended ascii character (accented chars etc.) in the 8-bit ascii range also. You would need something like tr[\x20-\x7e\x80-\xff][]cd; to avoid this and retain those.

thanks for this really detailed reply.
parvesh

s/\W+/ /g;  # replace non-word characters with a space
s/[^a-zA-Z0-9_.*$]+//g;  # Delete characters that aren't word characte
+rs, periods, asterisks, or dollar signs.
[download]