[Perl] my quiz (easy)

[Mikhael Goikhman]

On 23 Oct 2002 11:47:50 +0200, Shlomo Yona wrote:
> 
> Write a sub which receives some integer number 
> greater than zero (lets call the number N), and
> return a string of N characters (over [A-Za-z0-9]).
> 
> Calling the sub within the same run of the program should 
> guarantee no two strings of length N are equal (i.e. all the
> strings of same length are uniqe).

Here is the solution with the best performance (both memory and speed).
It circulates: "000" .. "999", "AAA" .. "ZZZ", "aaa" .. "zzz", "000".
It utilizes a feature of the ++ operator for "\d+" and "\w+" strings.

#!/usr/bin/perl -w
use strict;

my %uniques = ();
sub unique {
	my $len = shift;
	# let's limit our unique string length from 1 to 99 chars
	return undef unless defined $len && $len =~ /^[1-9]\d?$/;

	# initialize to "000" if this is the first time with $len
	$uniques{$len} ||= "0" x $len;

	# advance to the next unique string in one of the 3 ranges
	return $uniques{$len}++ if length($uniques{$len}) == $len;
	
	# otherwise switch the range: digits -> upper -> lower -> digits
	my $overflow = substr($uniques{$len}, 0, 1);
	$uniques{$len} = "A" x $len if $overflow eq "1";
	$uniques{$len} = "a" x $len if $overflow eq "A";
	$uniques{$len} = "0" x $len if $overflow eq "a";
	return $uniques{$len}++;
}

# test this function
printf "Expected string: '%s', got string: '%s'\n", $_, unique(1)
	foreach ("0" .. "9", "A" .. "Z", "a" .. "z", "0" .. "9", "A");

printf "Expected string: '%s', got string: '%s'\n", $_, unique(2)
	foreach ("00" .. "99", "AA" .. "ZZ", "aa" .. "zz", "00");

> Let's see who writes the nicest and perhaps smartest peace of code.

I am glad you don't ask for the "shortest". Although nice solutions are
usually short, the number of the chars used has nothing to do with this.

Regards,
Mikhael.