slow hashes with recursive loops?

[Hans de Groot <ha...@interspective.nl>]

Hi, 

I made a function to sort a folder list (with childs/parents), I know it must be a silly method I use but it was the only one I could think of at the time. 
The problem I now have is that is is very slow. I only have 128 folders but it takes like 5 secs to loop through this.

I have narrowed it down to the function below and I do not see the resaon why it would be sooooooooooooooo slow. specially if you considure I have many many hits on the database
and other silly this in the rest of the code. (It's for a website tool).

Then I remebered I read something about a bug or problem with embperl 2.x and hashes. I cannot find it anymore but I use embperl 1.3b5 on redhat 7.2 (but same on redhat 6.1)

Is there any work arround for this or something else I can do to speed things up? Right now I use $udat to store the tree list (ie I store the hashes in a string via join and every hit I split the string again into hashes. 
It seems to me that this process shoudl take much longer jan just loop through that loop I made. But it's a hack and for now it works. 
But people without cookies still get the slow version (Imagine 6 frames from which 3 use navigation is a 15 secs freeze.. )

Help
:-)

here my silly loop (note that the digit function does not take noticable time)

sub get_list()
  {
  my($tempparent)=shift;
  my($step_size)=shift;
  my($step_type)=shift;
  my($subkey);
  my($key);  
  $glob_lev++;
  foreach $key (sort {$parentlist{$a} <=> $parentlist{$b}} keys %parentlist)
    {
    foreach $subkey (sort {$node{$a} <=> $node{$b}} keys %node)
      {
      if (($key eq $tempparent) && (exists $level{$tempparent,$subkey})) # deze node heeft child.
        {    
        if ($glob_lev != $meet_lev)
          {  
          $meet_lev=$glob_lev;
          }  
        $count++;
        $spacer=""; 
        # Silly hash, 3 separate hashes for name, id, and level
        $ruk=&digit($count,6);

        $levelhash{$subkey}=$glob_lev;
        $treehash{$ruk}=$spacer.$nodenaam{$subkey}; #Voor naam met indent
        $parenthash{$ruk}=$key; 
        $key_hash{$ruk}=$subkey; # om de id terug te vinden

        &get_list($subkey,$step_size,$step_type);
        }    
      } # End for subkey 
    } #end for key 
  $glob_lev--;
  }


Regards

Hans


------
Hans de Groot
Interspective Internet Services B.V.
Tel: +31 (0) 30 6001180
Fax +31 (0) 30 6044204
www:   http://www.interspective.nl
email:   hansg@interspective.nl




---------------------------------------------------------------------
To unsubscribe, e-mail: embperl-unsubscribe@perl.apache.org
For additional commands, e-mail: embperl-help@perl.apache.org

Re: slow hashes with recursive loops?

[Gerald Richter <ri...@ecos.de>]

Hi,
>
> I made a function to sort a folder list (with childs/parents), I know it
must be a silly method I use but it was the only one I could think of at the
time.
> The problem I now have is that is is very slow. I only have 128 folders
but it takes like 5 secs to loop through this.
>
> I have narrowed it down to the function below and I do not see the resaon
why it would be sooooooooooooooo slow.

Also I don't quite understand what you are doing here, it seems that you are
doing many many time the same sort, while the aruments to sort doesn't
change. At least I can't see in your code where you change %parentlist or
%node. So the first thing to do is to move these sort out of the loop and
perform them _before_ the loop, save the result in a array and use this as
argument to foreach. This should speed up your code many times (unless I
have overseen something). Maybe there are more constant operation that could
be moved out of the loop. Since I have to less knowlegde of your
datastructure, I can't say if there are more optimaziation possible.

> Then I remebered I read something about a bug or problem with embperl 2.x
and hashes. I cannot find it anymore but I use embperl 1.3b5 on redhat 7.2
(but same on redhat 6.1)
>

That has nothing to do with Perl hash. That was about some hashs that
Embperl 2 uses internaly and that only occurs if you have very huge
iteration of the same html code.

BTW. If you are using 1.3b5 I would recommend to update 1.3.4, but that's up
to you and has nothing to do with you problem.

Gerald

-------------------------------------------------------------
Gerald Richter    ecos electronic communication services gmbh
Internetconnect * Webserver/-design/-datenbanken * Consulting

Post:       Tulpenstrasse 5         D-55276 Dienheim b. Mainz
E-Mail:     richter@ecos.de         Voice:    +49 6133 925131
WWW:        http://www.ecos.de      Fax:      +49 6133 925152
-------------------------------------------------------------



---------------------------------------------------------------------
To unsubscribe, e-mail: embperl-unsubscribe@perl.apache.org
For additional commands, e-mail: embperl-help@perl.apache.org