#!/usr/bin/perl

# benz - detects superset/subset relations between csv files
# by hoge1e3
#
# $0 filespec [ filespec ... ]
#    filespec :=  fileName[:colNum:[colNum ..]]
#
#  fileName can be *.csv(comma separated) or *.txt (tab separated).
#   colNum is either number or A-Z, corresponding column name on Excel worksheets.
#
#  example:
#     ./benz.pl students.csv:B score.csv
#  output:
#     students.csv_score.csv_.txt

use strict;
my %sets;
my $setid=1;
my $outFile="";
for my $a(@ARGV) {
  my ($f, @cols) = split(/:/,$a);
  if (@cols==0) { @cols=(0); }
  elsif (@cols==1 and $cols[0]=~/[A-Z]/ ) {@cols= &toInt($cols[0]); }
  my $set= bless {setid=>$setid, file=>$f , cols=>[@cols], colCount=>0 } ;
  $sets{$setid}=$set;
  $setid *= 2;
  $outFile.=$f."_";
}
$outFile .=".txt";
$outFile =~ s#/#__#g;

my %records;  # recid -> setid -> [@rec]
for my $setid(keys %sets) {
  my $set=$sets{$setid};
  $set->read;
}
open OUT,">$outFile" or die "Cannot open $outFile";
for my $recid(sort keys %records) {
  my $rec = $records{$recid};
  my $bits=0;
  my @buf;
  for my $setid(sort keys %sets) {
     my $set = $sets{$setid};
     $bits |= $set->{setid} if ($rec->{$setid}) ;
     push @buf, join ("\t", &pad( $set->{colCount}, @{$rec->{$setid}} ) ) ;
  }
  print OUT "$bits\t".join("\t",@buf)."\n";
}
close OUT;



sub read {
  my $t=shift;
  my $csv;
  $csv=1 if ($t->{file} =~ /csv$/);
  open IN ,$t->{file} or die("Not found $t->{file}");
  while (<IN>) {
     s/\n//;
     s/\r//;
     my @rec;
     if ($csv) {@rec=&parseExcelLine($_);}
     else {@rec=split (/\t/);}
     if (@rec>$t->{colCount}) {$t->{colCount}=@rec;}
     my @ids;
     for my $col(@{$t->{cols}}) {
        push @ids, $rec[$col];
     }
     my $recid=join("\t",@ids);
     my $rec=$records{$recid};
     if (not $rec) {
        $rec={};
        $records{$recid}=$rec;
     }
     if ($rec->{$t->{setid}}) {print STDERR "duplicate: $t->{file} - $recid\n"; }
     $rec->{$t->{setid}}= [@rec] ;
  }
  close IN;
}

sub pad {
   my ($num, @rec)=@_;
   while (@rec<$num) {
     push @rec,"";
   }
   @rec;
}

sub parseExcelLine {
    my $l=shift;
    $l =~ s/\n//;
    $l =~ s/\r//;
    $l .= ",";
    my @res=();
    while ($l) {
	if ($l =~ s/^""// ) {
	    push @res,"";
	} elsif ($l =~ s/^"// ) {
	    if ($l =~ /([^"])",/) {
		my $it=$`.$1;
		$l = ",$'";
		$it =~ s/""/"/g;
		#$it =~ s/"(?=")//g;
		push @res,$it;
	    } else {
		print STDOUT "excelerror : unterminated $l\n";
	    }
	} else {
	    $l =~ s/^[^,]+//;
	    push @res,$&;
	}
	$l =~ s/^,//;
    }
    @res;
}

sub toInt {
   my $cols=shift;

   my @res;
   my @cola=unpack("C*",$cols);
   for my $a (@cola) {
     push @res, $a-65;
   }
   @res;
}