How do I use map to merge an array of hashes? - perl

Question

Hoping someone could explain the proper process for this. I have a data structure like this:
[
{
Author => "101",
WK1 => "",
WK10 => "",
WK11 => "",
WK12 => "",
WK13 => "",
WK14 => "X",
WK15 => "",
},
{
Author => "102",
WK1 => "",
WK10 => "",
WK11 => "",
WK12 => "X",
WK13 => "X",
WK14 => "",
WK15 => "",
WK2 => "X",
WK3 => "X",
WK4 => "",
},
{
Author => "101",
WK1 => "",
WK10 => "",
WK11 => "",
WK12 => "",
WK13 => "",
WK14 => "X",
WK15 => "X",
},
]
WK keys may or may not exist. The Author key will always exist, and I’m trying to merge those duplicates. Overwriting values is ok (if defined).
Should I try map (not sure how), or would it be better to create a new hash using the Author as key, pushing the other values into an array? Something like
foreach $x (#$data) {
%new_hash = ...
...
}
scratching head

Answer 1

It sounds like you are starting with something like the following:
my $records = [
{ Author => "101", WK1 => "", WK2 => "X", WK3 => "X" },
{ Author => "101", WK1 => "X", WK2 => "", WK4 => "X" },
{ Author => "102", WK1 => "", WK2 => "", WK3 => "X" },
];
It sounds like you want to produce the following:
my $records_by_author = (
"101" => { Author => "101", WK1 => "X", WK2 => "", WK3 => "X", WK4 => "X" },
"102" => { Author => "102", WK1 => "", WK2 => "", WK3 => "X" },
);
The following will do that:
my %records_by_author;
for my $record (#$records) {
my $author = $record->{Author};
$records_by_author{$author} = {
( $records_by_author{$author} ? %{ $records_by_author{$author} } : () ),
%$record,
};
}
If instead you want the following output:
my $records = [
{ Author => "101", WK1 => "X", WK2 => "", WK3 => "X", WK4 => "X" },
{ Author => "102", WK1 => "", WK2 => "", WK3 => "X" },
];
Just add the following:
$records = [ values(%records_by_author) ];
It you want this output and you want to preserve the original order, let me know.

Answer 2

I suggest merging along the lines of the code below. From each anonymous hash in #$data, grab the author identifier and then update the merged WK values where we never overwrite a non-empty string with an empty one.
You could do the same with map, and that would be a worthwhile exercise. It seems to read much more naturally as nested structural loops.
my %authorwk;
for (my $i = 0; #$data; ++$i) {
local $_ = shift #$data;
die "$0: missing Author in record $i" unless exists $_->{Author};
my $author = $_->{Author};
while (my($wk,$present) = each %$_) {
next unless $wk =~ /^WK/;
$authorwk{$author}{$wk} = $present
if $present || !exists $authorwk{$author}{$wk};
}
}

Answer 3

map is inappropriate here: a series of slice assignments is all that is necessary
It looks like this. Note that the only duplicate author 101 has an identical set of keys in both instances, so I've added key WK7 to the first instance and WK8 to the second so that you can see that they are both added to the result
use strict;
use warnings 'all';
my $data = [
{
Author => 101,
WK7 => "7", WK1 => "", WK10 => "", WK11 => "",
WK12 => "", WK13 => "", WK14 => "X", WK15 => "",
},
{
Author => 102,
WK1 => "", WK10 => "", WK11 => "", WK12 => "X",
WK13 => "X", WK14 => "", WK15 => "", WK2 => "X",
WK3 => "X", WK4 => "",
},
{
Author => 101,
WK8 => "8", WK1 => "", WK10 => "", WK11 => "",
WK12 => "", WK13 => "", WK14 => "X", WK15 => "X",
},
];
my %authors;
for my $item ( #$data ) {
my $author = $item->{Author};
#{ $authors{$author} }{keys %$item} = values %$item;
}
use Data::Dump;
dd \%authors;
output
{
101 => {
Author => 101,
WK1 => "",
WK10 => "",
WK11 => "",
WK12 => "",
WK13 => "",
WK14 => "X",
WK15 => "X",
WK7 => 7,
WK8 => 8,
},
102 => {
Author => 102,
WK1 => "",
WK10 => "",
WK11 => "",
WK12 => "X",
WK13 => "X",
WK14 => "",
WK15 => "",
WK2 => "X",
WK3 => "X",
WK4 => "",
},
}