[EMAIL PROTECTED] wrote:
> Good Afternoon
>
> I am attempting to develop a script that will parse a directory
> listing and return only directory names that match a given
> expression.
>
> It would make sense to me to use File::Find to do this but based on
> the dir structure I am parsing, the amount of overhead to do this is
> enourmous !
>
> Basically, I have a file structure similar to:
>
> Dir1\Dir2\Support\119404\dirx\diry
> Dir1\Dir3\Support\119893\dirx
> Dir1\Dir4\Support\188884\dirx\diry\dirz
> .....
> Dir1\Dir1000\Support\100858
>
> I am simply interested in finding the directories directley under the
> Support dir (ex.119404 from the 1st example) . There is no
> consistancy to the naming convention other then Dir1 and Support.
> Dir2 can be many different values.
>
> I tried functionality similar to the following that did work on a
> much smaller test bed:
>
> my $dirs="I:\\ID_000000_000999";
> find sub { push @dirs, $File::Find::dir if $File::Find::dir =~
> m/.+[Ss]upport\/\d+$/;}, $dirs;
>
> But in a larger scale dir structure. The performance of this was
> horrible !! (Since it is looking through the entire structure
> including dirs under the directory I am trying to match on).
>
> As you can see form the I:\\ this is on windows, so ls and similar
> UNIX commands are not available.
>
> Any thoughts on how I can accomplish this task with the lowest amount
> of overhead?
>
> Thanks!
> Jason
Hi Jason,
Did you look into globs? You need to make a list of every directory
you're interested in ONCE, then process that list.
This worked for me:
# I don't have access to a windows system with Perl
# installed, but as this is all native perl, it
# should be fine
my $dirs='/ID_000000_000999';
my @dirs_found;
push @dirs_found, $_ while <$dirs/*/[Ss]upport/*>;
You can check out these perldocs for further info:
perldoc -f glob
perldoc File::Glob
perldoc perlop (look for I/O Operators)
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
