If you are using a small hash do not bother about the load If you are going to use huge hashes I think you should create subhashes or sorted arrays on the fly using *map* . I personally am not very fluent with map I think you will get better help on the mailling list with a more direct question
On Tue, 2002-09-17 at 19:58, Jeff AA wrote: > > > -----Original Message----- > > From: Ramprasad A Padmanabhan [mailto:[EMAIL PROTECTED]] > > Sent: 17 September 2002 14:44 > > To: [EMAIL PROTECTED]; Jeff Aa > > Cc: [EMAIL PROTECTED] > > Subject: Re: sorting a hash - multiple key fields > > > > > > Use this . I think you wud do better later to move the sort > > function to > > a seperate sub > > > > foreach $key (sort { > > ($$h{$a}{taste} eq $$h{$b}{taste} ) ? $$h{$a}{name} cmp > > $$h{$b}{name}: $$h{$a}{taste} \cmp $$h{$b}{taste} > > } (keys %$h)) { > > print "$key \n"; > > } > > > > > > Did you use the string comparisons 'eq' and 'cmp' deliberately on the > numeric values? Is this faster than a numeric compare? > > A little explanation on why you think the ? operator is more [efficient > | faster | better | clearer | other ] would be appreciated. > > You seem to be recommending returning a sorted list of keys, rather than > a sorted list of sub-hashes, and you are doing hash lookups inside the > sort function, using $a and $b as key values - surely this would be > slower than having $a and $b already loaded with hashrefs? > > TIA > Jeff > -- Ramprasad A Padmanabhan Sr Software Engineer Netcore Solns Pvt Ltd Mumbai ph - (022) 4628000 -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
