TL;DR

Itâ€™s Perl Weekly Challenge time again, now on issue #080 TASK #1.

And by the wayâ€¦ do you know that depending on the time zone you might already start submitting pull requests for the HacktoberfestÂ®?!?

# The challenge

You are given unsorted list of integers @N. Write a script to find out the smallest positive number missing.

# The questions

There are not too many questions that can be asked, butâ€¦

• the list contains integers, but the question asks for a numberâ€¦ well, letâ€™s assume that it must be an integer!
• is it correct to assume that the answer to an empty list is `1`?
• how to deal with invalid inputs?

# The solution

To be honest, nothing clever came to mind. Which is probably good, especially if I put myself in the shoes of someone doing a job interview!

So my plan is pretty boring:

• sort the list
• go through it and find the lowest positive integer thatâ€™s missing from it.

With one little twistâ€¦ why sort the whole list? Negative integers donâ€™t really matter, so we will filter them out beforehand.

After much talking, hereâ€™s the core of the solution:

``````1 sub spnb {
2    my @Np = (0, sort(grep { \$_ > 0 } @_));
3    push @Np, \$Np[-1] + 2;
4    ((\$Np[\$_] + 1 < \$Np[\$_ + 1]) && return \$Np[\$_] + 1) for 0 .. \$#Np - 1;
5 }
``````

Line 2 filters out non-positive input numbers and then sorts them. I also force the addition of a `0` at the very beginning, it will not alter the sorting of the whole thing and will allow me to simplify looking through the array.

Line 3 adds another element that does not break the sorting and lets me treat the last element as any other one (i.e. as an â€śinternalâ€ť element).

Line 4 is a concession to some golfing, letâ€™s take a closer look to its alternative unrolled and readable form:

``````4.1 for my \$index (0 .. \$#Np - 1) {
4.2    my \$candidate = \$Np[\$index] + 1;
4.3    if (\$candidate < \$Np[\$index + 1]) {
4.4       return \$candidate;
4.5    }
4.6 }
``````

We iterate over all but the last element, which was added by us. At each loop, we compute what would be the candidate value, i.e. our possible return value, which would be one more than the element we are currently analyzing (line 4.2).

If this candidate is less than the next element in the array (line 4.3), then itâ€™s indeed missing and we can return it (line 4.4).

If you want to cut-and-paste, hereâ€™s a complete file for doing some experimentation:

``````#!/usr/bin/env perl
use 5.024;
use warnings;
use English qw< -no_match_vars >;
use experimental qw< postderef >;
no warnings qw< experimental::postderef >;

sub spnb {
my @Np = (0, sort(grep { \$_ > 0 } @_));
push @Np, \$Np[-1] + 2;
((\$Np[\$_] + 1 < \$Np[\$_ + 1]) && return \$Np[\$_] + 1) for 0 .. \$#Np - 1;
}

for my \$test (
[ 5, 2, -2, 0 ],
[ 1, 8, -1 ],
[2, 0, -1 ],
[],
[1, 2, 3],
) {
say spnb(\$test->@*);
}
``````

Cheers!

Comments? Octodon, , GitHub, Reddit, or drop me a line!