On Fri, 5 Mar 2004 19:18:04 -0500, Erik Hatcher <[EMAIL PROTECTED]>
wrote:
Thanks for the idea for a good example for the upcoming Lucene in Action
book... it's been added!
Thanks for mentioning me in the book ;)
Well, I actually already had a comment in the book about why you'd override getRangeQuery, and it said this:
* handle number ranges by padding to match how numbers were indexed
You did give me the incentive to flesh this out into an example.
I also created a variant of this to parse range queries like this field:[1/1/04 TO 12/31/04] into YYYYMMDD syntax so it becomes field:[20040101 TO 20041231]. This is very handy when dealing with dates in a typically more sensible YYYYMMDD format and allowing users to deal with them naturally also.
What about boolean fields? It's certainly not a good idea to use "true" or
"false" strings...
What about them? It all depends on how you want users to be able to query based on that flag. Do you want them to say field:true? field:on? field:yes? How you translate things in QueryParser is up to you - and this may of course have some impact on how you index. You could use "0" and "1" instead, and do the translation in a QueryParser subclass if you like.
BTW, isn't it slow to treat everything as strings?
Ummm, yeah.... Lucene is real slow! :)
You tell us.... is it slow with your data and environment? If so, give us some more details on the scenario.
Erik
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
