Re: [MacPerl-AnyPerl] About This .pm Stuff

[Richard Gordon <maccgi@bellsouth.net>]

At 07:02 -0400 5/2/1999, Chris Nandor wrote:
>It depends on what your data looks like.  For data where you have key/value
>pairs separated by a |, you can do:
>
>  %hash = map {split /\|/, $_, 2} <FILE>;
>
>If you have a line without a pipe, though, this will give you problems,
>because only one element from that line will go into the hash, and the next
>key will become the previous key's value instead of becoming a new key.  It
>all depends on your data.

That's what I was looking for and I used a \t to delimit keys, so it 
should work.

>Well, you can't have a space between O_ and RDWR, and you should not put
>".dbm" on the filename.  Some DBM libraries will put an extension on for
>you, some won't.  Just leave it off and you'll be better off.  Also, the
>tie statement is wrong.  You need:
>
>  tie VARIABLE, CLASSNAME, LIST
>
>So:
>
>  tie(%thehash, 'AnyDBM_File', $filename, $flags, $permissions) or die $!;

I typed instead of pasted and there isn't actually a space between O_ 
and RDWR in the script, but you have a good eye in catching that in 
the post. Are you saying that the statement should reference 
variables for flags and permissions? I was going by some stuff in the 
Camel book in doing it as I did.

>
>Let us know if that doesn't work.  Also, are you using -w and use strict?
>Further, are you checking for errors, as my example did?

I wasn't using strict but had -w on. If you mean using "or die" for 
error checking, I am using || die "Couldn't tie hash: $!;"

>Bart gave good suggestions for question 3, and I don't think I have
>anything else to add to it.

As I indicated in reply to him, changing the keys to lower case may 
not be an option in this particular case. The only reason that I was 
interested in using a tied hash was because the file is pretty big 
and I figured that something that was disk based might be a little 
more efficient if it meant that the whole file was not read into 
memory every time that the script ran. However, it may be just as 
fast and more flexible to simply open the file and look for a match 
from a line start. I guess my question is why bother with a hash, 
tied or otherwise?

Thanks for your help.


Richard Gordon
--------------------
Gordon Consulting & Design
Database Design/Scripting Languages
mailto://maccgi@bellsouth.net
770.565.8267

==== Want to unsubscribe from this list?
==== Send mail with body "unsubscribe" to macperl-anyperl-request@macperl.org