Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!usc!ucsd!ogicse!intelhf.hf.intel.com!agora!markb
From: markb@agora.uucp (Mark Biggar)
Newsgroups: comp.lang.perl
Subject: Re: Uniq, anyone?
Message-ID: <1990Oct13.182400.22738@agora.uucp>
Date: 13 Oct 90 18:24:00 GMT
References: <1990Oct12.075314.15054@utgard.uucp> <JGREELY.90Oct12142916@morganucodon.cis.ohio-state.edu>
Distribution: na
Organization: Open Communication Forum
Lines: 30

In article <JGREELY.90Oct12142916@morganucodon.cis.ohio-state.edu> J Greely <jgreely@cis.ohio-state.edu> writes:
>In article <1990Oct12.075314.15054@utgard.uucp> chris@utgard.uucp
> (Chris Anderson) writes:
>>On occasion, I find myself wanting something on the order of:
>>	@foo = (1,1,2,2,3,3);
>>	@bar = &uniq (@foo);
>
>This seems to do it, although it strips null fields out entirely (I
>call that a feature, myself):
>
>sub uniq {
>	local($o);
>	grep(($_ ne $o) && defined($o = $_),sort @_);
>}

This requires sorting the list before uniq'ing it, which make the algorithm
nlogn.  The following algorithm is better:

sub uniq {
	local(%item) = ();
	for (@_) {
		$item{$_} = 1;
	}
	keys(%item);
}

This produces an unsorted result that can of course can be sorted.
--
Perl's Maternal Uncle
Mark Biggar