Hi,

I have something that I don't understand and I thought I would throw it out here and see if someone can make sense of it.

I had a Lucene index, generated and updated by Hibernate Search. I just emptied my database, added 5 records, purged the index, reindexed and optimized (using Hibernate) using the Massindexer. When I open the index directory with Luke it shows me an index with 5 records. Fine. However, when I query the index with another Lucene based tool, it gives me back the results based on the old index. Sure enough, the index directory contains a larger cfs file that is older than the other files in the index directory. It looks like purging left an artifact for some reason.

Is there a reason and older index file might still linger?

hello,
the fact that two different tools show different results can only mean that one of them is wrong ;)
after a purgeAll() documents are virtually deleted, to really reclaim disk space and remove the no-longer used file segments you might want to optimize() the index.

I've no clue about why this older tool might be able to read the should-have-ignored segment - that's a bug of the other tool, possibly it's using an old incompatible lucene version.