[MarkLogic Dev General] xpath string construction
rob at koberg.com
Tue Oct 14 12:20:27 PDT 2008
OK, I jumped in on this thread late without reading the previous
posts. I just did and see what you guys are talking about -- I was
thinking of something else.
We just lock the document for editing (if necessary, unlocking when
someone's session ends). Relying on unique IDs at the element level
would have problems as well (rare, but...). If an element is deleted,
it's unique ID is fair game to be used for a new element since it is
still unique. I suppose you could store all IDs ever used in a
document and check it :(
Use small documents and lock them :)
Sorry for the noise.
On Oct 14, 2008, at 2:36 PM, Wayne Feick wrote:
> Hi Robert,
> The problem is that the document is read in one transaction to
> generate a web page, and then updated in a later transaction when
> you hit the submit button on that web page.
> On Tue, 2008-10-14 at 14:29 -0400, Robert Koberg wrote:
>> What I don't get is why come up with a unique identifier scheme when
>> you already have URIs. If the problem is trying to avoid the same doc
>> getting update from 2 different directions at the same time, then
>> wouldn't ML's transaction management be the solution?
> General mailing list
> General at developer.marklogic.com
More information about the General