lucene-solr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Otis Gospodnetic <otis_gospodne...@yahoo.com>
Subject Re: trie fields default in example schema
Date Mon, 03 Aug 2009 02:41:04 GMT
Would it make sense to instead add new tint(eger) type instead of renaming integer to pinteger?
(thinking about people upgrading to Solr 1.4).

Otis
--
Sematext is hiring -- http://sematext.com/about/jobs.html?mls
Lucene, Solr, Nutch, Katta, Hadoop, HBase, UIMA, NLP, NER, IR




----- Original Message ----
> From: Yonik Seeley <yonik@lucidimagination.com>
> To: solr-dev@lucene.apache.org
> Sent: Sunday, August 2, 2009 3:01:09 PM
> Subject: trie fields default in example schema
> 
> I'm working on a jumbo trie patch (just many smaller trie related
> issues at once) - SOLR-1288.
> 
> Anyway, I think support will be good enough for 1.4 that we should
> make types like "integer" in the example schema be based on the trie
> fields.  The current "integer" fields should be renamed to "pinteger"
> (for plain integer), and have a recommended use only for compatibility
> with other/older indexes.  People have mistakenly used the plain
> integer in the past based on the name, so I think we should fix the
> naming.
> 
> The trie based fields should have lower memory footprint in the
> fieldcache and are faster for a lookup (the only reason to use plain
> ints in the past)... sint uses StringIndex for historical reasons - we
> had no other option... we could upgrade the existing sint fields, but
> it wouldn't be quite 100% compatible and there's little reason since
> we have the trie fields now.
> 
> -Yonik
> http://www.lucidimagination.com


Mime
View raw message