From: dhtapp <dhtapp@...>
Date: 2004-02-25T07:34:50+09:00
Subject: Re: HTML/XML Parsing...

My first suggestion would be to
1.  make a to-many relationship to the image records,
2.  to store any IMG attributes parsed from the HTML in the image records
themselves (with maybe an ordering attribute within the to-many set, in case
sliced images are bumped up against each other for positioning), and
3.  To decide whether to store the images themselves on a filesystem with
pathnames in the records, or to store image data as BLOBS within the records
themselves.

If you need to retrieve the images through a non-HTTP pipeline into another
process, then BLOBS may be the way to go.  If I was simply going to generate
dynamic HTML, then I'd probably go ahead and put the images out on a
filesystem where both the database and the Webserver could get to 'em.

- dan




"Ruby Tuesday" <rubytuzdayz@yahoo.com> wrote in message
news:c1gidb$1j6abi$1@ID-205437.news.uni-berlin.de...
> I'm wondering if anyone ever come across an example on how to parse an
html
> table(with images) using either XSLT or Ruby scripts.
>
> I'd like to be able to extract all the data and put them in the
> database(MySQL, SQLite, etc).
>
> There's a twist though, some of the image cells has 2 or more jpeg images
> instead of one. Since I'm not an expert database designer, how do you do
> that?
>
> Table fields:
> ------------
> xnum    text(40)         unique
> image   jpeg image     (may have none, or 1+ images)
> desc     memo(256)
> loc       memo(256)
>
> Thanks.
>
>
>