On Sat, 30 Jun 2001 01:20:51 +0200, allan wrote:
>my (main) question is wheter it is worthwhile to create a txt-file of
>the relevant data on each line and then build a hash based on this file.
>this hash would probably never have more than a few hundred elements. my
>aiming is partly to unstress the database server but primarily to
>speeden things up as it is concerning a heavily visited website.
I would not bother with creating a separate database export file. In my
experience, it's easy to forget to update the export file, and you're
doing you're thing with outdated data.
But loading data into a hash isn't a bad idea. I've done that with
thousands of items. But I'd just do one SELECT statement (if it's an SQL
db server), and get them all at once, and store them into the hash
directly.
BTW using "<TAG ID=1>" as the hash key isn't really smart. For one, a
hash is case sensitive. Since all the tags are "TAG", using just the ID
as the hash key sounds less like asking for trouble. Oh, and
s/$tag/$id/ig does sound like a bad idea, too. You know what: you can do
your processing in the RHS of the s///, with an /e modifier. So
> while (/(<TAG[^>]+id=(\d+)>)/ig) {
> my $tag = $1;
> my $id = $2;
> my $value = getValue($id);
> s/$tag/$value/ig; #replace
> }
can become
s{(<TAG[^>]+id=(\d+)>)}{
my $tag = $1;
my $id = $2;
my $value = getValue($id);
$value
}ieg;
(and that's still more verbose than necessary) where getValue() is
essentially a hash lookup.
--
Bart.
