transliterating.. char-conversation

raptor has asked for the wisdom of the Perl Monks concerning the following question:

hi, I had to convert several symbols from German alphabet to different represantation... and used something like this until now :

sub blah { $_ = shift; s/ä/ae/g; s/Ä/AE/g; s/ö/oe/g; s/Ö/OE/g; s/ü/ue/
+g; s/Ü/UE/g; s/ß/ss/g; $_ }
[download]

But I now need more complicated conversation so I started rewriting(see code below)... but after I thougt a litle bit I decided first to ask, may be there is easier way... these language depenend stuff has always been a black-box to me.. ...reading pelrlocale and POSIX to clear out the things a little bit... may be u can help me..
This just from the scratch I even not tested it .. on the NT it seems that dosen't work.. (going to Linux console :") )

$tstr = 'abc äÄöÖüÜß test';

%subsTable = (
    'ä' => 'ae', 'Ä' => 'AE', 'ö' => 'oe', 'Ö' => 'OE',
    'ü' => 'ue', 'Ü' => 'UE', 'ß' => 'ss'
);
$notInSQL = qr{'|`|´|"|,|%|$};
$inSQL = qr{`|´|"|,|$};

sub blah { 
    my $str = shift; 
    my $regex = join "|", map quotemeta, keys %subsHash;
    print $regex,"\n";
    $regex = qr/$regex/;
    $str =~ s/($regex)/$subsHash{$1}/g;
#    if ($str =~ /^\s*(SELECT|INSERT|UPDATE|DELETE)/i ) { $str =~ s/$i
+nSQL/ /g }
#        else { $str =~ s/$notInSQL/ /g };
#may be have to use tr/// here        
    return $str
};


print "orig : $tstr\n";
print blah($tstr)."\n";
[download]

these are not all symbols to be converted I'm expecting more to come...

Comment on transliterating.. char-conversation Select or Download Code

Replies are listed 'Best First'.
(Ovid) Re: transliterating.. char-conversation by Ovid (Cardinal) on Aug 08, 2001 at 00:32 UTC
Without looking at your code too closely, I see some areas for improvement. As a good rule of thumb, don't alternate on single characters. By adding strict (and switching the alternation to character classes for a performance boost), I got your script to work: use strict; my $tstr = 'abc äÄöÖüÜß test'; my %subsTable = ( 'ä' => 'ae', 'Ä' => 'AE', 'ö' => 'oe', 'Ö' => 'OE', 'ü' => 'ue', 'Ü' => 'UE', 'ß' => 'ss' ); my $notInSQL = qr{['`´",%\$]}; my $inSQL = qr{[`´",\$]}; sub blah { my $str = shift; my $regex = join "", map { quotemeta $_ } keys %subsTable; $regex = "[$regex]"; print "Regex: $regex\n"; $str =~ s/($regex)/$subsTable{$1}/g; return $str }; print "orig : $tstr\n"; print blah($tstr)."\n"; [download] When I added strict, I immediately discovered that you had referred to `%subsTable` as `%subsHash` in your subroutine. By switching the variable name, it worked fine. Cheers, Ovid Vote for paco! Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.	[reply] [d/l]
Re: (Ovid) Re: transliterating.. char-conversation by raptor (Sexton) on Aug 08, 2001 at 00:48 UTC
thanx ... i'm doing this in write-and-run script and later all will be me moved in module where I'm always using 'strit' of course :").. but see it is needed even in write-and-run scripts thanx for char-classes I forgot about them :")	[reply]
Re: transliterating.. char-conversation by raptor (Sexton) on Aug 08, 2001 at 00:27 UTC
oops my mistake : it has to be %subsHash instead of %subsTable....:") but never mind a better solutions are always welcome... also some discusion about locales-and-friends...	[reply]