Ari Halberstadt wrote:
> Raphael Finkel <raphael@cs.uky.edu> wrote:
> >Is PostgreSQL appropriate for a database of documents? I see several
> >potential problems.
> >
> >1. Tuples are limited to 8KB. I assume even the "text" type is limited that
> >way. But my documents may be longer than that.
>
> This would pretty much nix the idea of using postgres for a document
> repository. You'd need to split the data into separate rows in another
> table. I've done this for my bulletin board (though most posts should be
> less than 8K in my case), but I think it would be even more of a hastle
> with documents. What you can do instead is store descriptive information in
> the DB and store the actual documents in the file system with a field in
> the document table containing the path to the file. A few years ago I
> worked on a commercial document management program that stored attributes
> in an Oracle database but kept the actual documents on disk.
>
That a good solution, but now that XML is starting to break off it kinda puts
PostgreSQL users at a disadvantage. We have to have static XML documents on on
disk in order to use the format and have external search engines to search for
content in the documents. It is not that big of a headache, but is an
inconvenience. I will be doing some work on XML soon for my job and they are
going straight Oracle because they want to create "searchable" XML documents on
the fly to post and distribute on the web. Very LARGE documents. Like the legal
code for NYC and such. When Postgre starts using large objects it is going to
make a lot of noise. Especially for "start-up" companies.
--
Mark Fleming Student ICQ# 16171200
e-mail: mtony@iname.com
http://www.globaldrum.com/mfleming
http://coolbox.bowiestate.edu/~mfleming
