There's a section on doing fast joins using range indexes in the "Inside 
MarkLogic Server" paper.

http://www.marklogic.com/resources/inside-marklogic-server/

-jh-

P.S.  Why so anonymous?  We don't bite.

On May 23, 2013, at 11:33 PM, [email protected] <[email protected]> 
wrote:

> Hi,
> 
> I have two sets of documents in ML as:
> 
> fn:doc()/a. (a is root element)
> Sample structure:
> <a>
> <b>"sample1"</b>
> </a>
> 
> And fn:doc()/c. (c is root element)
> Sample structure:
> <c>
> <b>"sample2"</b>
> </c>
> 
> Now I have millions of both these types of docs.
> I have to join these two sets of docs on the element b.
> Return the matching values.
> 
> Currently I am using nested for loop as:
> 
> for $doc1 in fn:doc()/a
> for $doc2 in fn:doc()/c
> where ($doc1/b = $doc2/b)
> return 
> (
> <output>
> {
> <b>{doc1/b}</b>
> }
> </output>
> )
> 
> This takes very long even for a few thousands. I am trying to join them using 
> cts:search. Any help will be appreciated.
> 
> Thanks..
> 
> _______________________________________________
> General mailing list
> [email protected]
> http://developer.marklogic.com/mailman/listinfo/general

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

Reply via email to