I currently have this example test case working against lucene RAMDirectory.....
very simple test case... https://github.com/deanhiller/nosqlORM/blob/memoryindexer/input/javasrc/com/alvazan/test/TestIndexAndOrParens.java with this Entity and his namedQueries.... https://github.com/deanhiller/nosqlORM/blob/memoryindexer/input/javasrc/com/alvazan/test/db/Activity.java We allow applications to split their 1 billion/1 trillion, whatever into 1 million or 1 billion or whatever number of indexes and dba's can query their partitions by grabbing the index and running standard SQL. Currently, yes, we translate from SQL to lucene query language. This has really just been a playground and now I would like to start doing joins. 1. I was wondering though, does SOLR only support a join of two indexes or can you join more than two indexes? 2. Also, anyone know if solandra has support for this? Currently, we actually cache all writes and only send them in the flush so if any additions to the index fail earlier nothing was written so there is a better chance of data being consistent. thanks, Dean