What do you mean by "unicode characters"? This is a multiple-choice type question. Possible choices are: UTF-16BE (aka UCS-2BE), UTF-16LE (aka UCS-2LE), utf8. (If you don't understand the question, you need to learn enough about unicode so that you do, or at least know enough about your input data to know which choice is correct.)

You should be using a very recent version of Perl (5.8.1 or later), and you should look up the man pages "perluniintro", "perlunicode" and the "Encode" module.

To read unicode data properly from a file or STDIN, (and to write it properly to a file or to STDOUT), the easiest way is to use the appropriate character-encoding "IO Layer" (see docs for PerlIO). Something like this:

my $open_mode = "<:utf8"; # or "<:encoding(UTF-16LE)" etc.
open( IN, $open_mode, "unicode_input.txt" );

binmode( STDOUT, ":utf8" ); # or whatever form of unicode is supported
+ by your display tool

while (<IN>) {  # data will be read as (or converted to) utf8 on input
   # do stuff with $_, then
   print;
}
[download]

Apart from that, the code you posted has a lot of unnecessary effort and temporary storage. You read into a "prehasha" array, then copy that into "char" and "pinyin" arrays, then finally try to put that into a hash (though it's not clear that you succeed there). You could just read into the hash. Following on the snippet above (where the file is being read as unicode data):

my %charpinyin;

while (<IN>) {
   chomp;
   my ( $chchar, $pinyin ) = split /\t/;
   $charpinyin{$chchar} = $pinyin;
}
[download]

(updated last snippet to reflect the OP's use of tab-delimited data)