It's pretty easy to create a situation where count(doc($uri)) => 2. Whether or 
not that's a bug depends on how the database got into that state. If you have a 
single forest with duplicate documents, and you haven't been doing anything 
tricky to the forest filesystem, that would be a pretty serious bug.

But the more usual case is that two different forests each have a fragment with 
the same document URI. That wouldn't necessarily indicate a server bug. It 
could be a bug, but in my experience is more likely to be a result of user 
error. A developer or user might have switched around some forest assignments. 
URIs are guaranteed to be dataabse-unique during normal database transactions, 
but admin forest-assignment operations carry no such guarantee.

-- Mike 

On 26 Jun 2013, at 14:24 , Florent Georges <[email protected]> wrote:

> Michael Blakeley wrote:
> 
>> Well, isn't the most likely explanation that the backup also
>> contains the duplicate?
> 
>   Yes, actually that's my point, the backup contains the duplicates
> (which was not necessarily the case), so we can reproduce it, then
> maybe can investigate further :-)
> 
>> Either way you should contact support if you want to dig into the
>> problem more.  If there's a bug involved, they will want to know
>> about it.
> 
>   Having fn:count(fn:doc()) returning more than 1 is definitely a bug.
> Yes, we did contact support and opened a ticket since.
> 
>   Thanks!  Regards,
> 
> -- 
> Florent Georges
> http://fgeorges.org/
> http://h2oconsulting.be/
> 

_______________________________________________
General mailing list
[email protected]
http://developer.marklogic.com/mailman/listinfo/general

Reply via email to