Re: Bidirectional lookup algorithm? (Updated: further info.)

Since your strings and integers are unique - why don't you use a hash for strings and an array for integers

#!/usr/bin/perl -l

my %str2int;
my @int2str;

$i=0;
for ('aaaa'..'zzzz') {
    $str2int{ $_ } = ++$i;
    $int2str[$i] = $_;
}
[download]

and look up the strings by index into the array, and the integers by key into the hash?

perl -le'print map{pack c,($-++?1:13)+ord}split//,ESEL'

Comment on Re: Bidirectional lookup algorithm? (Updated: further info.) Download Code

Replies are listed 'Best First'.
Re^2: Bidirectional lookup algorithm? (Updated: further info.) by Laurent_R (Canon) on Jan 05, 2015 at 19:00 UTC
The integers are 64-bits. I do not know what they really look like, but there is clear danger of ending up with a sparse array growing to be much much larger than the hash.	[reply]
Re^3: Bidirectional lookup algorithm? (Updated: further info.) by shmem (Chancellor) on Jan 05, 2015 at 22:34 UTC
Right you are. qwurx [shmem] ~/comp/perl/tmp > perl -le '@a = 2*64; for(1..29){$a[2 +$_]=$_}print `ps -o vsz= -p $$`' Out of memory! [download] Having 8 GB RAM. 2*28 doesn't die. perl -le'print map{pack c,($-++?1:13)+ord}split//,ESEL'	[reply] [d/l]