Not that I know of. And, if you create & index such Lucene documents, and then do a MatchAllDocsQuery sorting by your field, this is (unfortunately) not an external sort! Ie, Lucene loads all terms data in RAM as packed byte[], for merging the per-segment results.
It even does this, unnecessarily, for an optimized segment, even though we only need ords in that case (there's an issue open for this). Doing a sort-by-String-field without loading the String data even when there are multiple segments in the index would be a nice addition :) Mike McCandless http://blog.mikemccandless.com On Tue, Jun 14, 2011 at 6:31 AM, Dawid Weiss <[email protected]> wrote: > Hi. While I was playing with automata recently, I had a use case > scenario when I could really use an external sort of a large list of > unicode strings. I know I could simply emulate this by creating > synthetic documents, index, etc., but is there a more "direct" way of > achieving this using Lucene's internals? > > Dawid > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
