random behaviour of perl hashes

evilgoblin has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: random behaviour of perl hashes by lima1 (Curate) on Aug 28, 2006 at 00:16 UTC
from keys: The keys are returned in an apparently random order. The actual random order is subject to change in future versions of perl, but it is guaranteed to be the same order as either the "values" or "each" function produces (given that the hash has not been modified). Since Perl 5.8.1 the ordering is different even between different runs of Perl for security reasons (see "Algorithmic Complexity Attacks" in perlsec).	[reply]
Re: random behaviour of perl hashes by Tanktalus (Canon) on Aug 28, 2006 at 00:39 UTC
Generally speaking, even if Perl were to give you the same output each time, I doubt it would be sage advice to count on what that order was in any way. Merely upgrading perl could give you a different set of answers (if Perl-5-Porters were to suddenly find a new magical algorithm that solved some issue of speed, memory, or both, for example). Which means that if you moved from one machine (running Perl 5.8.1) to another (running Perl 5.8.3, say), you could get different answers. Thus, the answer has always been that if you need some sort of order to these keys, you need to impose that order yourself (via sort or Tie::IxHash or something). As lima1 says, they now force this good habit on you to reduce security exposure risks. But that doesn't change the fact that it was always a good idea.	[reply]
Re^2: random behaviour of perl hashes by evilgoblin (Novice) on Aug 28, 2006 at 01:08 UTC
ah but this thread suggests differently atleast for 'small' hashes. By small I mean with less than 20 keys in a bucket. Edit g0n: converted href to pm link	[reply]
Re: random behaviour of perl hashes by blue_cowdawg (Monsignor) on Aug 28, 2006 at 02:19 UTC
basically is there a random component, every time the program is run, to the ordering of the hash keys given that the key and the values stay the same I'm a firm believer in doing empirical testing so I wrote a little sniglet of code: `#!/usr/bin/perl -w use strict; use Data::Dumper; my @ingredients = ( [ qw / onions diced / ], [ qw / potatos diced / ], [ qw / paprika spicy / ], [ qw / salt shaken / ], [ qw / pepper fresh / ] ); my $hash_browns={}; foreach my $ingredient (@ingredients){ $hash_browns -> { $ingredient -> [0] } = $ingredient -> [1]; } foreach my $key ( keys %$hash_browns ){ printf "%s\t%s\n",$key,$hash_browns->{$key}; }` [download] When I run this code three times I get something like this: `[pberghol@cowdawg hash]$ perl hash.pl salt shaken onions diced pepper fresh paprika spicy potatos diced [pberghol@cowdawg hash]$ perl hash.pl salt shaken onions diced pepper fresh paprika spicy potatos diced [pberghol@cowdawg hash]$ perl hash.pl salt shaken onions diced pepper fresh paprika spicy potatos diced` [download] Looks pretty much the same to me. Now, I'm not sure why this is all important, but if I were looking for random behavior, I'd program for it and not count on a "freak of nature" in my programming language of choice to give it to me. On the other hand if predictable behavior is what I want, again I'll program for that. Peter L. Berghold -- Unix Professional Peter -at- Berghold -dot- Net; AOL IM redcowdawg Yahoo IM: blue_cowdawg	[reply] [d/l] [select]
Re^2: random behaviour of perl hashes by gellyfish (Monsignor) on Aug 28, 2006 at 08:00 UTC
As lima1 pointed out this behaviour has been deliberately changed as of the perl 5.8.*, I get this with 5.8.1: `[jonathan@orpheus jonathan]$ perl foo.pl potatos diced pepper fresh salt shaken paprika spicy onions diced [jonathan@orpheus jonathan]$ perl foo.pl pepper fresh potatos diced onions diced paprika spicy salt shaken [jonathan@orpheus jonathan]$ perl foo.pl salt shaken paprika spicy onions diced potatos diced pepper fresh` [download] but yes it would be a mistake to rely on an any hash ordering behaviour in a programme. /J\	[reply] [d/l]
Re^3: random behaviour of perl hashes by swampyankee (Parson) on Aug 28, 2006 at 20:48 UTC
I get this running ActiveState's 5.8.8 (build 817): `Case 1 salt shaken onions diced pepper fresh paprika spicy potatos diced Case 2 salt shaken onions diced pepper fresh paprika spicy potatos diced Case 3 salt shaken onions diced pepper fresh paprika spicy potatos diced` [download] I suspect that it has to do with the details of the implementation of the hashing algorithm, which may differ from system to system. My interpretation of the "random ordering" in the perldocs is simply that "we refuse to guarantee the order of the keys for a hash". emc Only two things are infinite, the universe and human stupidity, and I'm not sure about the former. Albert Einstein	[reply] [d/l]
Re^4: random behaviour of perl hashes by liz (Monsignor) on Aug 28, 2006 at 20:57 UTC
Re^5: random behaviour of perl hashes by swampyankee (Parson) on Aug 28, 2006 at 21:13 UTC
Re: random behaviour of perl hashes by Joost (Canon) on Aug 28, 2006 at 21:39 UTC
During one process, the order given by each, keys and values will stay the same as long as the hash doesn't change. The order is not guaranteed to be the same if you start a new interpreter - i.e. running a perl program twice will probly not give the same order. "What should it profit a man, if he should win a flame war, yet lose his cool?"	[reply]
Re^2: random behaviour of perl hashes (rarely) by tye (Sage) on Aug 28, 2006 at 21:49 UTC
i.e. running a perl program twice will probly not give the same order. s/probly not/almost always/, actually. But the "almost" is the important part. (: Don't rely on the order not changing, but you'll be even more disappointed if you try to rely on the order not staying the same. - tye	[reply]


Welcome to the Monastery
	PerlMonks