On Wed, Nov 4, 2009 at 4:09 PM, Christine Schwartz
<[log in to unmask]> wrote:

> > Does MarkLogic operate directly on the XML or does it index it?

> Mark Logic is a native XML database and most of the code is XQuery. So, it's
> operating directly on the XML. It has specific optimized queries that it
> runs.

That's what their marketing material says, but in fact it does index
the full-text content like a traditional search engine. You can also
run pure XQuery without setting up indexes, but you'll be limited in
the scope of queries available and the performance you'll get from
unindexed searches.

I agree that if it's possible, combining the files at the time of
injection into ML makes sense.  (On some of the projects I've worked
on with ML that would've been infeasible due to the scale of the data
and the frequency of updates.)