On Thu, Jun 20, 2013 at 7:09 PM, Tom Burton-West <tburt...@umich.edu> wrote: > Sine my objective is to undersdand BlockGroupingCollector, I thought I > would extract a test for BlockGroupingCollector, maybe in another file as > Robert suggested.
I think that makes sense! > I was looking at testRandom because that is where BlockGroupingCollector is > tested, but it looks to me like some trickery is involved creating the > block join type index near the comment where it says" // Build 2nd index, > where docs are added in blocks by group, so we can use single pass > collector " > > In testBasic, I can see the call to indexWriter.addDocuments(), but I don't > see that happening anywhere in testRandom. It's in the confusingly named getDocBlocksReader method. That method takes the GroupDoc array and uses IW.addDocuments to add each group as a block of documents. > Should I just be looking at testBasic for a simple/understandable way to > create a block-join index for later testing of the BlockGroupingCollector? Hmm but testBasic seems not to test BlockGroupingCollector. Maybe the first step is to write testBasic in your new TestBlockGrouping... > Also I'm a bit confused by the use of Doc Values. Is that necessary to use > the BlockGroupingCollector, or are they in there in order to test some > feature related to Doc Values and the Collectors? Not necessary; doc values are mixed into that test because we can do non-block grouping by doc values. Mike McCandless http://blog.mikemccandless.com --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
