Distinct-N | Compute Distinct-N metric proposed by Jiwei Li et al | Hashing library

So it isn't quite clear how you want your newly "denormalized" nodes to be composed, but from the example output given I am going to assume you want to join all distinct values and separate them by a whitespace.

The following should do the job:

I think you can filter using where condition as below:

If you're using JBossEAP 7/WildFly 10.x then this is JavaEE 7, although the same answer applies to Java EE 6.

You only appear to be using Local interfaces, so none of the instructions that you linked apply because they are only for remote EJB clients. Therefore these statements:

In this context, WF10 will automatically create a JNDI mapping for this EJB as:

There is no standard XQuery function fn:distinct-nodes(...), XQuery only knows fn:distinct-values(...).

The third-party XQuery function library functx knows a functx:dinstinct-nodes(...) function, which is again implemented as standard XQuery functions. The library can be downloaded and imported as a module with most XQuery implementations, as it only uses standard XQuery functions.

If all the elements contains is the author's name, consider applying fn:distinct-values(...) instead and recreate the elements afterwards.

For performance reasons, it might be reasonable to only extract the required functions if compile time increases too much (the library is rather large). Also be aware that some of the functions have much faster XQuery 3.0 counterparts, taking advantage of new language capabilities.

fn is the default XQuery function namespace, funct the one defined by the function library.