Re: Read and write UTF-8

[Norah]

Here is an example:

#!/usr/bin/perl
use Encode qw/encode decode/;
open (INFILE, "< :encoding(UTF-8)", "utf8.txt") || die "blah blah blah
+";
open (OUTFILE, "> :encoding(UTF-8)", "oututf8.txt") || die "blah blah"
+;
while (<INFILE>)
{
  $line = $_;
  chomp ($line);
  $linestart = substr($line,0,20);
  $outline = "First 20:  "."$linestart";
  print OUTFILE "$outline\n";
}
close (INFILE);

Actually this one reads and writes the non-ASCII characters, but when there is a non-ASCII character in the record it doesn't count the correct # of characters.

[Corion]

You talk about characters - when using UTF-8, length and substr count characters, not octets, so in the output, you can find more than 21 octets. If you were already talking about characters, not octets, can you please show some short example input that exhibits the problem, preferrably together with a hexdump of the relevant portion of the file?

[choroba]

If you want to extract a given number of bytes from a UTF-8 string, use bytes::substr :

#!/usr/bin/perl
use warnings;
use strict;
use feature qw{ say };

use open IO => ':encoding(UTF-8)', ':std';

my $string = join q(), map chr, 9312 .. 9321;
say $string;
say substr $string, 0, 7;
say bytes::substr $string, 0, 7;

[hippo]

it doesn't count the correct # of characters.

With no sample data, that's pretty hard to verify. Have you read length() miscounting UTF8 characters? There's lots of good info in there (length works analagously to substr) along with easy examples of how to provide code complete with data which demonstrates the issue.

[Norah]

And to be clear - what it did here - it did the exact same thing when I removed the encoding from the open statements.

[Norah]

Here is simple input data:

Year*JEDocSrcP_USERE_DATE  P_DATE  CurLine
2011617 GJ448 Bruce12/20/1101/01/11USD1500
2011617 GJ349áBruce12/20/1101/01/11USD1500
2011617 GJ350 Bruce12/20/1101/01/11USD1500
2011617 GJ351 Bruce12/20/1101/01/11USD1500

The output looks like this:

First 20:  Year*JEDocSrcP_
First 20:  2011617 GJ448 Bruce1
First 20:  2011617 GJ349áBruce
First 20:  2011617 GJ350 Bruce1
First 20:  2011617 GJ351 Bruce1

Note that the asterisk * is really the UTF-8 heart symbol ♥. It wasn't displaying correctly here so I just put an asterisk there. I will work on the hex dump for you.

but I think it is counting octets

[Corion]

Note that the hexdump of your input data has a Byte Order Mark (BOM) at the front of it, which Perl counts at least as some characters.

Discounting the BOM, I get the expected output with the following program:

#!/usr/bin/perl -w
use strict;
use Encode qw/encode decode/;

open (INFILE, "<:encoding(UTF-8)", "utf8.txt") || die "blah blah blah"
+;
open (OUTFILE, ">:encoding(UTF-8)", "oututf8.txt") || die "blah blah";
binmode STDOUT, ':encoding(UTF-8)';

print "Ruler : [12345678901234567890]\n";
while (my $line = <INFILE>) {
  chomp ($line);
  print "Input : [$line]\n";
  my $linestart = substr($line,0,20);
  my $outline = $linestart;
  print "20    : [$outline]\n";
  print "---\n";
  print OUTFILE "$outline\n";
}
close (INFILE);

To remove the BOM at the start of your file, use maybe simply

$line =~ s!^\N{BYTE ORDER MARK}!!;

[Anonymous Monk]

Thank you for the suggestion - but I get the same results. I think there is some system setting on the laptop that will not read in characters - it just does bytes.