Shot in the dark: You're using Lucene to read the index. That's sort of circumventing all the typing stuff that Solr does. Solr can deal with an index where some of the segments are in one format (say 1.4) and others are in another (3.6). Maybe they're being stored in a format in the newer (or older) segments that doesn't work with raw retrieval of the values through Lucene in the same way.
Maybe it's able to retrieve the "stored" value from the indexed representation in one case rather than needing to store it. I'd query your index using EmbeddedSolrServer instead and see if that changes what you see. Michael Della Bitta Applications Developer o: +1 646 532 3062 | c: +1 917 477 7906 appinions inc. “The Science of Influence Marketing” 18 East 41st Street New York, NY 10017 t: @appinions <https://twitter.com/Appinions> | g+: plus.google.com/appinions w: appinions.com <http://www.appinions.com/> On Fri, Jun 14, 2013 at 4:15 PM, Mingfeng Yang <mfy...@wisewindow.com>wrote: > I have an index first built with solr1.4 and later upgraded to solr3.6, > which has 150million documents, and all docs have a datefield which are not > blank. (verified by solr query). > > I am using the following code snippet to retrieve > > import org.apache.lucene.index.IndexReader; > import org.apache.lucene.store.*; > import org.apache.lucene.document.*; > > IndexReader input = IndexReader.open(indexDir); > Document d = input.document(i); > int maxDoc = input.maxDoc(); > for (int i = 0; i < maxDoc; i++) { > System.out.println(d.get('date'); > } > > However, about 100 million docs give null for d.get('date') and about other > 50 million docs give the right values. > > What could be wrong? > > Ming- >
