How do I replace special character?

[laredotornado]

Hello,

I'm on Solaris and trying to use a Perl command to replace some
untypeable charcters in a text document. When I open the document in
"vi", I see a couple of strange characters -- "~Q" (a single character,
not the combination of "~" and "Q") and "~R". How do I find out the
hex codes for these characters, and then how do I use the following
command

perl -pi -e 's/pattern_to_search/pattern_to_replace/g' myfile.txt

to replace them?

Thanks, - Dave

 

[Guest]

(e-mail address removed) wrote:
: Hello,

: I'm on Solaris and trying to use a Perl command to replace some
: untypeable charcters in a text document. When I open the document in
: "vi", I see a couple of strange characters -- "~Q" (a single character,
: not the combination of "~" and "Q") and "~R". How do I find out the
: hex codes for these characters, and then how do I use the following
: command

Looks a bit like old IBM cp437 coded material opened in a Latin1 environ-
ment. If you have any incarnation of vim, placing your cursor over the
character in question; hitting 'g' 'a' will show you the ascii, hex and
octal values of the character under the cursor.

: perl -pi -e 's/pattern_to_search/pattern_to_replace/g' myfile.txt

Say \x{<number>} in pattern_to_search for each character you want to
replace. Obtain <number> with the vi(m) command just mentioned.

Oliver.

 

[Tad McClellan]

trying to use a Perl command to replace some
untypeable charcters in a text document.

Many years ago I wrote a little utility to expose "funny characters"
when I encountered a similar situation. (it is pretty slow though.)

Here it is.

-----------------------------
#!/usr/bin/perl
# list_control_chars - output the values of control chars
use strict;
use warnings;

while (<>) {
foreach my $char (split //) {
my $chnum = ord($char);

if ( ($chnum >= 32 && $chnum <= 126) ||
$chnum == 9 ||
$chnum == 10 ) {
# do nothing
}
else {
my $hex = sprintf '0x%X', $chnum;
my $oct = sprintf '0%o', $chnum;
print "$chnum $oct ($hex)\n";
}
}
}

 

[Tad McClellan]

On unix, use the od (octal dump) program to dump files by octal or hex

On any OS that has perl installed, you can use xdump to dump files by hex.

xdump was sample code given in the 1st edition of the Camel book (p272).


Google found it, of course, though from a different book:

http://www.cpan.org/scripts/nutshell/ch6/xdump

 

[Ilya Zakharevich]

[A complimentary Cc of this posting was sent to
Tad McClellan

Many years ago I wrote a little utility to expose "funny characters"
when I encountered a similar situation. (it is pretty slow though.)

Should be trivial to speed up.

while (<>) {
foreach my $char (split //) {
my $chnum = ord($char);

if ( ($chnum >= 32 && $chnum <= 126) ||
$chnum == 9 ||
$chnum == 10 ) {

13 is not funny too; but your code is not in binmode, so you should
not see it; however, then on \cZ + Dosish you get unexpected stuff.

# do nothing
}
else {
my $hex = sprintf '0x%X', $chnum;
my $oct = sprintf '0%o', $chnum;
print "$chnum $oct ($hex)\n";
}
}
}

printf "%#07d %#07x %#07o\n", ($1) x 3 while /([^ -\x7E\r\n\t])/g;

Hope this helps,
Ilya