Re^5: create an xml file from column file

[hdb]

Another change in requirements? Wow. Full rewrite needed for this one.

use strict;
use warnings;
use Text::CSV;

my $csv = Text::CSV->new( { sep_char => "\t" } ); # assuming tab separ
+ated input
open my $words, "<", "words.txt" or die "Cannot open words.txt: $!\n";

# the following shows what's in words.txt; it is NOT words.txt itself!
=head words.txt
how            o                B-NP
are            o                I-NP
you            o                I-NP
some        o                o
really    o                B-GP 
=cut

my $lasttype;
my @text;
while( my $row = $csv->getline( $words ) ) {
        my $text = $$row[0];
        my $type = ( $$row[-1] =~ /\w-(\w+)/ ) ? $1 : "";
        $lasttype = $type unless @text; # special treatment for first 
+row
        if( $type eq $lasttype ) {
                push @text, $text;
        } else {
                print '<key="type">'."$lasttype</key><text>@text</text
+>\n" if $lasttype;
                $lasttype = $type;
                @text = ( $text );
        }
}

# print what's left over when all input read
print '<key="type">'."$lasttype</key><text>@text</text>\n" if $lasttyp
+e;

close $words;

[lakssreedhar]

is it possible to do without CSV .The code i wrote is below.But the required output is not coming

$out=" ";
my @text;
while(<>)
{
chomp;
@arr=split(/\s+/,$_);
push @r,$arr[0];
push @e,join' ',($arr[0],$arr[2]);
}

print '<text>';
print join ' ',@r;
print '</text>';
print"\n";

for my $a (@e) {
@q=split(/\s/,$a);
if($q[1]=~/^B-/)
{
$p=$';
$m=$q[0] if( $q[1] =~/^B-/ or $q[1]="o");
$out.="<annotation>\n<key=type>$p</type>\n<text>$m</text>\n</annotatio
+n>\n";
}
}
print"$out\n";

[hdb]

At your own risk, splitting on tab:

use strict;
use warnings;

open my $words, "<", "words.txt" or die "Cannot open words.txt: $!\n";

my $lasttype;
my @text;
while( <$words> ) {
        chomp;
        my @row = split /\t/;
        my $text = $row[0];
        my $type = ( $row[-1] =~ /\w-(\w+)/ ) ? $1 : "";
        $lasttype = $type unless @text; # special treatment for first 
+row
        if( $type eq $lasttype ) {
                push @text, $text;
        } else {
                print '<key="type">'."$lasttype</key><text>@text</text
+>\n" if $lasttype;
                $lasttype = $type;
                @text = ( $text );
        }
}

# print what's left over when all input read
print '<key="type">'."$lasttype</key><text>@text</text>\n" if $lasttyp
+e;

close $words;

[lakssreedhar]

It is working but this error is popping up Use of uninitialized value in pattern match (m//) at f.pl line 14, <$words> line 18.