On 20/01/2015 23:12, Flavel Heyman wrote: > I’m trying to create indexes in Jackrabbit Oak (1.1.3) with MongoDB. I was > following http://jackrabbit.apache.org/oak/docs/query/query.html to create > indexes. > > The problem is that I’m not sure what to do with the NodeBuilder after it is > generated... It feels like I need to commit/merge the “indexNodeBuilder”, > but that hasn’t worked for me yet. If I try to do an “ns.merge(builder, > commitHook, info)” it fails because the NodeBuilder generated by the > IndexUtils is not a DocumentRootBuilder as required by DocumentNodeStore > (that I am using with MongoDB). > > > Any Suggestions on how to load the indexes created in Java to the JCR? I > have fullTextSearch running, but it requires me to > setFullTextComparisonWithoutIndex to true, which means my indexing isn’t > working. Otherwise if it is false it won’t return any results in my search. > The purpose of this is to prototype what it would look like to switch from > Jackrabbit 2.8 to Oak for speeding up searches. > I'm not understanding what you're trying to do. Are you trying to define a new index definition, let's say a property index on `jcr:createdBy` or create a new index implementation?
If you're trying to create a new index definition, what's the query you're trying to optimise? To define a new index I would suggest to create the node structure using the JCR api rather than the NodeBuilder. HTH Davide
