I found a very cool presentation. Yep it is about XML, The concepts behind XQuery to be particular, but it is easy relating to it with a (Oracle) relational mindset.
Its about choosing the right storage (which doesn’t have to physical), about the language, which index will or won’t suffice your needs, access paths, compression, and for instance, how to optimize data retrieval. The first part is all about XQuery, the “second” part is also interesting for the Oracle (performance) minded people (even if you hate XML in general). It also gives you an idea off what the relational world since 1970 has accomplished and had to deal with (and still does).
XML In its nature, is more demanding than relational data (IMHO), mainly due its free format. A lot is still undiscovered country regarding a proper solution. Some ideas have already settled in my mindset like the statement on slide 96: “There is no one fits all solution” and some of my other “solutions” I thought about and wanted to investigate further, in this case for structured XML, are old, very very old…(slide 135, 1999: “XML Schema to Relational Mapping”)
Another cool thing is that I found it. It will give me a lot to think about and a better understanding of query optimizers and data handling in general.
- Have a nice read: “XML Query Processing”: Daniela Florescu, Donald Kossmann
Wasn’t BEA bought by Oracle? Didn’t Ms Florescu once work for Oracle?
Anyway. Cool stuff.
😎
I believe this is an old-ish slideset, put together by Dana when she was at BEA before she moved to Oracle.
If you ever want a simple, scalable XML and XQuery platform and are prepared to move away from the relational universe and into the native XML universe, you might consider trying MarkLogic Server.
Alternative link for the one above: http://165.246.104.120/icde/crawled/icde/2004/Florescu-Kossman.pdf