What I am looking for is some guidance to Perl and Databases. I am in the process of designing perl scripts which will need to be able to read a small file (preferably a perticular database format) for instructions, but will then link to another database which has the potential to be huge in size.

I was thinking along the lines of DBM Files or XML. MySQL had crossed my mind, but at the end of the day, it would need to be able to used on many machines, and MySQL is not installed on every machine. XML is more likely, but is it really designed for database side of things?

Any help?

i ve almost never worked with xml but read about it. there is xml query language to query data stored in xml similar to sql. so i guess it's what u need. but i don't know how fast it handles huge piles of data.
is the program intended to run on *nix or various platforms? there's no perl on doze by default either, is it?
cheers, jens

I got no experience whatsoever with XML, but if your files
tend to get huge performance might become an issue, and
you'd probably wish for a SQL database engine

PostgreSQL is quite portable and not very huge ... :}

Cheers,
Tink

Yeah thanks for the info!

It is intended solely for Linux, which I believe that most have perl installed as default (and if not, why not! ).

To have a size guestimate, I would say approx 500 - 1000 rows, which would contain 5 -7 fields. What do you think something like RPM's use as their database format?

Gidday!

In that case you can probably even use a flat file format, without
indexing or a hash, since not everybody has LibXML or the
Perl XML module(s) installed, either :} and Perl is fast enough for
doing files with that few lines decently.

Cheers,
Tink

Thanks. I did think that a flat file format would suffice, and your right, the XMLmodule would need to be installed on the other machines.

I think I will start off with a basic file format, and if performance drops considerably then I guess I'll have to change. Cheers for all the info!

Sounds like a good idea ... worst case scenario would be to get (I bet there's something out there :})
some kind of indexing algorithm in Perl and throw that in when performance drops ... that way you'd
stay library independent ... and as it's not multi-user (is it?) you don't have to consider stuff like
record locking ... ;)

Wasn't a lot of info, just a few basics ... and it was my pleasure :}

Cheers,
Tink