Lists Home |
Date Index |
> I have a simple question, "Why do we need updates as part of
> the XQuery language at all?"
I asked the same question when I first got interested in XML databases; I
had assumed that document replacement would meet most of the requirement. I
still think that's true for many applications. But there are some where that
doesn't work at all well, for example a system that maintains a log of
events, with a frequent need to append data relating to new events.
In fact there are many XML database applications where the "document" is a
very natural structuring unit, and there are some where it's not a natural
unit at all: an example I use of the latter is the database for a national
railway timetable, where deciding what goes together in one document is a
key (and not easy) design decision. Knowing the unit of update is one of the
inputs to this decision.