"Also please do not recommend using map in void context."

In general, I agree with this statement.

"This is considered bad practice by many programmers since you are only making the call for its side-effects whilst throwing away its output."

However, I disagree with this reason.

The issue with "throwing away its output" was resolved, about 14 years ago^[perlhist], in Perl 5.8.1. From "perl581delta: Miscellaneous Enhancements":

"map in void context is no longer expensive. map is now context aware, and will not construct a list if called in void context."

The reason I would give for not using map in void context is based on efficiency. I've had reason to benchmark map vs. for on a number of occasions over the years: I've always found for to be measurably faster than map; and, unsurprisingly, map EXPR is faster than map BLOCK.

Here's a very quick and dirty benchmark I just wrote to show this:

#!/usr/bin/env perl

use strict;
use warnings;

use Benchmark 'cmpthese';

my @x = 0 .. 99;

cmpthese 1e6 => {
    map_expr => sub { my @y = @x; map ++$_, @y; return },
    map_block => sub { my @y = @x; map { ++$_ } @y; return },
    for => sub { my @y = @x; ++$_ for @y; return },
};
[download]

I ran this five times. All results were much the same. This one was roughly in the middle:

              Rate map_block  map_expr       for
map_block  75019/s        --      -44%      -47%
map_expr  134953/s       80%        --       -5%
for       141844/s       89%        5%        --
[download]

Benchmark run using:

$ perl -v | head -2 | tail -1
This is perl 5, version 26, subversion 0 (v5.26.0) built for darwin-th
+read-multi-2level
[download]