#!perl
use strict;
use Archive::Zip::MemberRead;
use XML::Twig;
use Text::CSV;

Archive::Zip::MemberRead->setLineEnd();
my $csv = Text::CSV->new ( {binary=>1, eol=>"\012"} )  
            or die "Cannot use CSV: ".Text::CSV->error_diag();

my $infile  = 'header1.xlsx';
my $outfile = 'results.csv';                 
open my $fh_out,'>',$outfile 
  or die "Could not open $outfile: $!";

my @sheet;
my $zip = Archive::Zip->new($infile);
for my $name ($zip->memberNames()){
  if ($name =~ /sheet(\d+).xml$/){
    $sheet[$1] = $name;
  }
}

my @fields = ('oddHeader','evenHeader','oddFooter','evenFooter');
$csv->print($fh_out, [$infile,scalar localtime()]);
$csv->print($fh_out, ['Sheet',@fields]);
my %text=();
for my $no (1..$#sheet) {
  %text=();
  process_sheet( $sheet[$no] );
  my @row = map{ $text{$_} } @fields;
  $csv->print($fh_out, [$no,@row]);
}
$fh_out->close or die "$!";
print "Results written to $outfile\n";

sub process_sheet {
  my $filename = shift;
  print "Extracting text from : $filename\n";
  my $fh_in  = Archive::Zip::MemberRead->new($zip, $filename);
  my $xml = $fh_in->getline();
  $fh_in->close();
  
  my $twig = XML::Twig->new( keep_spaces=>1,
    twig_roots => { 'headerFooter' => \&get_text },
  );
  $twig->parse( $xml );
}

sub get_text {
  my ($t,$elt) = @_;
  for ($elt->children){
    my $text = $_->text;
    # remove left center right format codes
    $text =~ s/&[LCR]/ /g;
    $text{$_->name} = $text ;
  }
}
[download]