perl compare elements of arrays and grouping - arrays

Question

I am back with another question. I have a list of data:
1 L DIELTQSPE H EVQLQESDAELVKPGASVKISCKASGYTFTDHE
2 L DIVLTQSPRVT H EVQLQQSGAELVKPGASIKDTY
3 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAN
4 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAG
5 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C LELDKWASL
6 L DIQMTQIPSSLSASLSIC H EVQLQQSGVEVKMSCKASGYTFTS
7 L SYELTQPPSVSVSPGSIT H QVQLVQSAKGSGYSFS P YNKRKAFYTTKNIIG
8 L SYELTQPPSVSVSPGRIT H EVQLVQSGAASGYSFS P NNTRKAFYATGDIIG
9 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
10 A MPIMGSSVVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
11 L DVVMTQTPLQ H EVKLDESVTVTSSTWPSQSITCNVAHPASSTKVDKKIE
12 A DIVMTQSPDAQYYSTPYSFGQGTKLEIKR
And I would like to compare the 3rd elements && 5th elements of each row, then group them if they have the same 3rd && 5th elements.
For example, with the data above, the results will be :
3: 3 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAN
4 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAG
5 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C LELDKWASL
9: 9 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
10 A MPIMGSSVVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
Fyi, in the actual data, the 3rd, 5th, 7th elements are very long. I have made them cut to see the whole.
This is what I have done, I know it is very clumsy, but as a beginner, I am doing my best.
And the problem is that it shows only the first set of 'same' group.
Could you show me where it went wrong and/or other pretty methods to solve this, please?
my $file = <>;
open(IN, $file)|| die "no $file: $!\n";
my #arr;
while (my $line=<IN>){
push #arr, [split (/\s+/, $line)] ;
}
close IN;
my (#temp1, #temp2,%hash1);
for (my $i=0;$i<=$#arr ;$i++) {
push #temp1, [$arr[$i][2], $arr[$i][4]];
for (my $j=$i+1;$j<=$#arr ;$j++) {
push #temp2, [$arr[$j][2], $arr[$j][4]];
if (($temp1[$i][0] eq $temp2[$j][0])&& ($temp1[$i][1] eq $temp2[$j][1])) {
push #{$hash1{$arr[$i][0]}}, $arr[$i], $arr[$j];
}
}
}
print Dumper \%hash1;

Answer 1

You appear to have overcomplicated this a bit more than it needs to be, but that's common for beginners. Think more about how you would do this manually:
Look at each line.
See whether the third and fifth fields are the same as the previous line.
If so, print them.
The looping and all that is completely unnecessary:
#!/usr/bin/env perl
use strict;
use warnings;
my ($previous_row, $third, $fifth) = ('') x 3;
while (<DATA>) {
my #fields = split;
if ($fields[2] eq $third && $fields[4] eq $fifth) {
print $previous_row if $previous_row;
print "\t$_";
$previous_row = '';
} else {
$previous_row = $fields[0] . "\t" . $_;
$third = $fields[2];
$fifth = $fields[4];
}
}
__DATA__
1 L DIELTQSPE H EVQLQESDAELVKPGASVKISCKASGYTFTDHE
2 L DIVLTQSPRVT H EVQLQQSGAELVKPGASIKDTY
3 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAN
4 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAG
5 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C LELDKWASL
6 L DIQMTQIPSSLSASLSIC H EVQLQQSGVEVKMSCKASGYTFTS
7 L SYELTQPPSVSVSPGSIT H QVQLVQSAKGSGYSFS P YNKRKAFYTTKNIIG
8 L SYELTQPPSVSVSPGRIT H EVQLVQSGAASGYSFS P NNTRKAFYATGDIIG
9 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
10 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
11 L DVVMTQTPLQ H EVKLDESVTVTSSTWPSQSITCNVAHPASSTKVDKKIE
12 A DIVMTQSPDAQYYSTPYSFGQGTKLEIKR
(Note that I changed line 10 slightly so that its third field will match line 9 in order to get the same groups in the output as specified.)
Edit: One line of code was duplicated by a copy/paste error.
Edit 2: In response to comments, here's a second version which doesn't assume that the lines which should be grouped are contiguous:
#!/usr/bin/env perl
use strict;
use warnings;
my #lines;
while (<DATA>) {
push #lines, [ $_, split ];
}
# Sort #lines based on third and fifth fields (alphabetically), then on
# first field/line number (numerically) when third and fifth fields match
#lines = sort {
$a->[3] cmp $b->[3] || $a->[5] cmp $b->[5] || $a->[1] <=> $b->[1]
} #lines;
my ($previous_row, $third, $fifth) = ('') x 3;
for (#lines) {
if ($_->[3] eq $third && $_->[5] eq $fifth) {
print $previous_row if $previous_row;
print "\t$_->[0]";
$previous_row = '';
} else {
$previous_row = $_->[1] . "\t" . $_->[0];
$third = $_->[3];
$fifth = $_->[5];
}
}
__DATA__
1 L DIELTQSPE H EVQLQESDAELVKPGASVKISCKASGYTFTDHE
3 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAN
2 L DIVLTQSPRVT H EVQLQQSGAELVKPGASIKDTY
5 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C LELDKWASL
7 L SYELTQPPSVSVSPGSIT H QVQLVQSAKGSGYSFS P YNKRKAFYTTKNIIG
6 L DIQMTQIPSSLSASLSIC H EVQLQQSGVEVKMSCKASGYTFTS
9 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
8 L SYELTQPPSVSVSPGRIT H EVQLVQSGAASGYSFS P NNTRKAFYATGDIIG
11 L DVVMTQTPLQ H EVKLDESVTVTSSTWPSQSITCNVAHPASSTKVDKKIE
10 A MPIMGSSVAVLAIL B DIVMTQSPTVTI C EVQLQQSGRGP
12 A DIVMTQSPDAQYYSTPYSFGQGTKLEIKR
4 A ALQLTQSPSSLSAS B RITLKESGPPLVKPTCS C ELDKWAG

Answer 2

Slightly different approach:
#!/usr/bin/perl
use strict;
use warnings;
my %lines; # hash with 3rd and 5th elements as key
my %first_line_per_group; # stores in which line a group appeared first
while(my $line = <>) {
# remove line break
chomp $line;
# retrieve elements form line
my #elements = split /\s+/, $line;
# ignore invalid lines
next if #elements < 5;
# build key from elements 3 and 5 (array 0-based!)
my $key = $elements[2] . " " . $elements[4];
if(! $lines{key}) {
$first_line_per_group{$key} = $elements[0];
}
push #{ $lines{$key} }, $line;
}
# output
for my $key (keys %lines) {
print $first_line_per_group{$key} . ":\n";
print " $_\n" for #{ $lines{$key} };
}

Answer 3

Example:
use strict;
use warnings;
{ ... }
open my $fh, '<', $file or die "can't open $file: $!";
my %hash;
# read and save it
while(my $line = <$fh>){
my #line = split /\s+/, $line;
my $key = $line[2] . ' ' . $line[4];
$hash{$key} ||= [];
push #{$hash{$key}}, $line;
}
# remove single elements
for my $key (keys %hash){
delete $hash{$key} if #{$hash{$key}} < 2;
}
print Dumper \%hash;

Answer 4

Your approach shows a pretty solid grasp of Perl idiom and has merit, but still is not how I would do it.
I think that you will have an easier time with this if you structure your data slightly differently: Let %hash1 be something like
(
'ALQLTQSPSSLSAS' => {
'RITLKESGPPLVKPTCS' => [3, 4, 5],
'ABCXYZ' => [93, 95, 96],
},
'MPIMGSSVAVLAIL' => {
'DIVMTQSPTVTI' => [9, 10],
},
)
where I have added a datum ABCXYZ which is not in your example to show the data structure in its fullness.

Answer 5

You should be using the 3-argument form of open() and you can simplify reading in the data:
open my $fh, '<', $file
or die "Cannot open '$file': $!\n";
chomp(my #rows = <$fh>);
#rows = map {[split]} #rows;
close $fh;
To group the rows, you can use a hash with the 3rd and 5th fields concatenated as the keys. Edit: You have to add a separation character to eliminate invalid results "if different lines produce the same concatenation" (Qtax). Additional data, for example, the number of the individual data rows, can be stored as the hash value. Here, the row's fields are stored:
my %groups;
for (#rows) {
push #{ $groups{$_->[2] . ' ' . $_->[4]} }, $_
if #$_ >= 4;
}
Sort out single elements:
#{ $groups{$_} } < 2 && delete $groups{$_}
for keys %groups;
greets,
Matthias