On Tue, May 12, 2009 at 2:30 AM, vivek sar <vivex...@gmail.com> wrote:
> Here is what I've read on maxMergeDocs, > > "While merging segments, Lucene will ensure that no segment with more > than maxMergeDocs is created." > > Wouldn't that mean that no index file should contain more than max > docs? I guess the index files could also just contain the index > information which is not limited by any property - is that true? > Yes, an individual segment will not contain more than maxMergeDocs number of documents. But the size of the segment may still vary because some documents may have more unique tokens than others. What you saw originally must have been a segment merge which is normal and happens in the course of indexing. I don't think there's a way to avoid that other than to have a ridiculously high mergeFactor (which will affect search performance). -- Regards, Shalin Shekhar Mangar.
