This implies there is no way to merge two parallel indexes(based on parallel reader) to get a new parallel index. Correct me if I am wrong.
On Tue, Jun 29, 2010 at 11:24 PM, Andrzej Bialecki <a...@getopt.org> wrote: > On 2010-06-30 05:12, Apoorv Sharma wrote: > > while calling addindexes or addindexes with no optimize can any gurantee > be > > given about the document order in the new documents given that the order > of > > directories/indexreader is fixed. > > > > So is it that ith document coming from jth indexreader will always have > some > > x(i,j) position in the final merged index ? > > Currently that should be the case. SegmentMerger copies contiguous > ranges of documents from readers (removing deletions), in the same order > as they were provided to addIndexes. Document numbers are shifted to > account for deletions, but other than that they are sequentially > increased according to each sub-reader's docBase+docNum. > > HOWEVER... this is an implementation detail, not any API guarantee, and > while it works today it may change silently in a minor Lucene revision. > Design your application so that it doesn't rely on this behavior. You > have been warned :) > > > -- > Best regards, > Andrzej Bialecki <>< > ___. ___ ___ ___ _ _ __________________________________ > [__ || __|__/|__||\/| Information Retrieval, Semantic Web > ___|||__|| \| || | Embedded Unix, System Integration > http://www.sigram.com Contact: info at sigram dot com > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org > For additional commands, e-mail: java-user-h...@lucene.apache.org > >
