"Peter Scott" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> On Thu, 12 May 2005 21:10:44 +0530, Madhur Kashyap wrote:
>
> > The chunk of perl code written below shows the way I have been using
> > the scalar referencing technique available in perl.
> >
> > $a="X15565/X123/35";
> > $b="n245";
> > ${$a."\t".$b}=0.598; ### <<== Scalar Referencing
> > ....
> > $x=tracename (1056.45,1076.56); ## returns X15565/X123/35
> > $y=tracename (234,34.89); ## returns n245
> > ${$x."\t".$y}+=0.63; ### makes life very simple
> >
> > For my application, the scalars held by $a and $b are actually some
> > net names. Typically, I will have, say around 300,000-400,000 such
> > names. Also, there is a physical quantity (floating number) which is a
<snip>
>
> I see no reason why you can't just use a hash lookup, which should be just
> as fast. I cannot tell from your code what property you are storing,
> which is also a red flag. Let's say that it is some interconnection cost:
>
> my %cost;
> [...]
> $cost{"$x\t$y"} += 0.63;
>
> See how more readable that is?
>
Or even simpler still, a multidemensional hash:
$cost{$x}{$y} += 0.63;
Todd W.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
