use warnings;
use strict;

use Data::Dump qw(dump);

my %db;             # Datenbank
my $using = 'V1';   # Spalte fuer JOIN

# Daten einlesen, hier aus DATA-Segment

for my $table ('table1', 'table2') {
    my $id_col;
    while (my $line = <DATA>) {
        last if $line !~ /\S/;

        # Spaltennamen incl. Reihenfolge sichern, 
        # Spaltennummer der $using-Spalte ermitteln
        
        if (! exists $db{'col_names'}{$table}) {
            my @cols = split ' ', $line; 
            grep { $id_col = $_ if $cols[$_] eq $using } 0..$#cols;
            die "$table does not contain col '$using'\n" if ! defined $id_col;
            @{$db{'col_names'}{$table}} = @cols;
            next;
	}

        my @cols = split ' ', $line;
        my $id = $cols[$id_col];

        # Daten in Datenbanktabelle einlesen
        $db{$table}{$id} = {
            map { @{$db{'col_names'}{$table}}[$_] => $cols[$_] } 0..$#cols
        };
    }
}

# Spaltennamen fuer Differenz-Tabellen und Join-Tabelle erzeugen,
# Namen, die nur in table2 vorhanden sind, an join-Spaltennamen anhaengen

@{$db{'col_names'}{'not_in_table1'}} = @{$db{'col_names'}{'table2'}};
@{$db{'col_names'}{'not_in_table2'}} = @{$db{'col_names'}{'table1'}};
@{$db{'col_names'}{'table_join'}}    = @{$db{'col_names'}{'table1'}};


for my $col (@{$db{'col_names'}{'table2'}}) {
    if (! grep { $_ eq $col } @{$db{'col_names'}{'table_join'}}) {
        push @{$db{'col_names'}{'table_join'}}, $col;
	}
}

# Daten mergen, Datensaetze aus table2 ueberschreiben die aus table1

for my $id (keys %{$db{'table2'}}) {
    if (exists $db{'table1'}{$id}) {
        %{$db{'table_join'}{$id}} = %{$db{'table2'}{$id}};
	}
    else {
        %{$db{'not_in_table1'}{$id}} = %{$db{'table2'}{$id}};
    }
}

for my $id (keys %{$db{'table1'}}) {
    if (! exists $db{'table2'}{$id}) {
        %{$db{'not_in_table2'}{$id}} = %{$db{'table1'}{$id}};
        %{$db{'table_join'}{$id}} = %{$db{'table1'}{$id}};
    }
}

# Entkommentieren, um Datenstruktur anzusehen:
# dump(%db);

# Ausgabe

no warnings 'uninitialized';

for my $table ('table_join', 'not_in_table1', 'not_in_table2') {
    printf(
        "$table: %d Zeilen\n\n%s\n",
        scalar(keys %{$db{$table}}),
        join("\t", @{$db{'col_names'}{$table}})
    );

    for my $row (sort keys %{$db{$table}}) {
        printf(
            "%s\n",
            join("\t",
                map { $db{$table}{$row}{$_} } @{$db{'col_names'}{$table}}
            )
        );
    }
    print "\n";
}

__DATA__
 V1 V2 V3 V4 V5
ID1 aa nn zz ii
ID2 ss rr oo ll
ID4 qq mm öö pp
ID5 nn vv bb xx

 V1 V2 V3 V4 V5 V6
ID1 aa nn zz ii 45
ID3 ff xx yy tt 45
ID4 qq mm öö pp 45
ID6 uu kk nn ff 45
ID7 qq ww ee rr 45