how to trim file - remove the rows which with the same value in the columns except the first two columns

Here I want to have your help on trimming a file, by remove the rows which with the same value in the columns except the first two columns.

the file I have (tab-delimited, with millions of rows, and tens of columns)

Jack Mike Jones Dan Was
1 2 7 3 4
2 3 9 4 8
T T C T T
T M T T T
W A S I S

the file I want (remove the rows which have the same values in cells except the first two)

Jack Mike Jones Dan Was
1 2 7 3 4
2 3 9 4 8
T T C T T
W A S I S

Could you give me any hints on my problem? Thanks a lot.

And I have experienced several e开发者_高级运维xcellent scripts of awk, shell and perl, in a related question. Thanks a lot for the helpers.

---

awk '{
    val=$3
    for (i=4; i<=NF; i++)
        if (val != $i) {
            print
            break
        }
}'

---

The simplest thing I could come up with (half joking:)

#!/usr/bin/perl
while (<>)
{
    my (undef, undef, @flds) = split;
    print if 1<scalar keys % {{ map { $_ => 1 } @flds }}
}

Explanation

_{It leverages a temporary hash table to find unique columns per line. Here goes:}

while (<>)   # for each line
{
    # split the line into columns, discarding the first two
    my (undef, undef, @flds) = split; 

    my %columns   = map { $_ => 1 } @flds; # insert the value as key into a hashtable
    my @uniq_cols = keys %columns;         # get just the keys
    my $uniq_count= scalar @uniq_cols;     # count the keys

    print if 1<$uniq_count                 # if count == 1, all columns are the same
}

To be even more explicit, the 'map' call is roughly equivalent to the usual idiom:

    # my %columns   = map { $_ => 1 } @flds;
    my %columns;

    foreach $fld (@flds)
    {
         $columns{$fld}++; # actually the map version does '$columns{$fld} = 1;' every time
    } 

HTH

---

Try this: perl -ne 'next if /^\w+\W+\w+\W+(\w+)(\W+\1)+\W*$/; print;'

That is, match:

^        beginning of line
\w+      first word
\W+      non-word (like spaces, tabs, etc)
\w+\W+   second word and spaces
(\w+)    third word (and remember)
(\W+\1)+ spaces followed by a copy of the third word as many times as necessary
\W*      optional trailing spaces
$        end of line