按键对Perl中的哈希数组排序

Question

I define an array of hashes like this:

while (my $line = <PILEUP>) {
    if ($line =~ /INDEL/ and $line !~ /^#/) {
        chop $line;
        my @splitline = split(/\t/, $line);
        $hash2{$splitline[1]}{len} = length($splitline[4]);
        $hash2{$splitline[1]}{type} = "INDEL";
        push @{$hash1{$splitline[0]}}, %hash2;
    }
}

Then, I want to sort that array by the key of each hash:

for my $chr (keys %hash1) {
    my @sorted =  sort { $a->{ } <=> $b->{ } } @{%hash1{$chr}};
}

However, no clue how to continue here.

Answer 1

I define an array of hashes like this:

No, you don't. You define a hash of arrays like that.

    chop $line;

chomp is preferable in just about every case. During 10 years of programming Perl, I've only ever used chop as a curiosity.

    $hash2{$splitline[1]}{len} = length($splitline[4]);
    $hash2{$splitline[1]}{type} = "INDEL";
    push @{$hash1{$splitline[0]}}, %hash2;

Here, you push the keys and values of %hash2 onto the array inside %hash1 . You probably wanted to push the hash ref: \\%hash2 . But in that case you need to make it lexically scoped, or else it will push the same reference multiple times.

for my $chr (keys %hash1) {
    my @sorted =  sort { $a->{ } <=> $b->{ } } @{%hash1{$chr}};
}

This is in the right direction, except that you have to choose one (or more) of the hash values to sort by. For example length and type. And you cannot use a loop.

my @sorted_keys = sort { $hash1{$a}{len}  <=> $hash1{$b}{len} ||
                         $hash1{$a}{type} cmp $hash1{$b}{type} } keys %hash1;