Hi, there are currently two projects, porting lucandra to HBase: http://github.com/akkumar/hbasene http://github.com/thkoch2001/lucehbase
hbasene stores a unique integer with each stored document, while lucehbase directly stores the user's primary key in the termVector table. Every lucehbase indexreader creates an internal map of integers to primary keys in ram. This means with lucehbase lucene will see new document ids with every new indexreader, while the document ids remain constant with hbasene. Do you see any problems with the approach of lucehbase? I found the following discussions, which seems to proof my point, that there isn't any problem: http://www.mail-archive.com/java-u...@lucene.apache.org/msg01665.html http://www.mail-archive.com/java-u...@lucene.apache.org/msg12172.html http://www.mail-archive.com/lucene-net-...@incubator.apache.org/msg00298.html http://www.mail-archive.com/lucene-...@jakarta.apache.org/msg06165.html Best regards, Thomas Koch, http://www.koch.ro --------------------------------------------------------------------- To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org For additional commands, e-mail: java-dev-h...@lucene.apache.org
