I found Lucene in Action perfect for 1.9 dev, but 2.0+ deviates a bit from the API used in "Lucene in Action".
The book is still immensely useful, but API differences may cause new readers some confusion. I believe "Lucene in Action" is written for 1.4 java package. Regards, -Vijay From: Simone Busoli [mailto:[EMAIL PROTECTED] Sent: Wednesday, 1 August 2007 7:40 AM To: [email protected] Subject: Re: test To everyone, please read Lucene in Action, it will teach you most of the things you need to know about Lucene. Simone Patrick Burrows wrote: Ah. that's got it. Thanks! On 7/31/07, feran <mailto:[EMAIL PROTECTED]> <[EMAIL PROTECTED]> wrote: Is the Field UN_TOKENIZED? If it's TOKENIZED you may not find it because the value will have been split into terms. ----- Original Message ----- From: "Patrick Burrows" <mailto:[EMAIL PROTECTED]> <[EMAIL PROTECTED]> To: <mailto:[email protected]> <[email protected]> Sent: Tuesday, July 31, 2007 3:21 PM Subject: Re: test Hm... I do this: Query q = new TermQuery(new Term("link", args[1])); (from a command line test app I made) and it still does not find the url that Luke is showing me in the index. args[1] has the exact url copied out of Luke. I get 0 hits back -- which is probably better than getting *every* document back... but still not ideal. On 7/31/07, Kurt Mackey <mailto:[EMAIL PROTECTED]> <[EMAIL PROTECTED]> wrote: Not only can you, but it's preferred. The QueryParser really only exists to handle human input. If you can do it programmatically, things are much easier. -----Original Message----- From: Patrick Burrows [mailto:[EMAIL PROTECTED] Sent: Tuesday, July 31, 2007 2:04 PM To: [email protected] Subject: Re: test Yeah. That's exactly what is happening. Didn't realize I could use my own query without going through the parser. On 7/31/07, Kurt Mackey <mailto:[EMAIL PROTECTED]> <[EMAIL PROTECTED]> wrote: Are you using the query parser thing for that? It will split on the various special characters in a URL, and (by default) give you something like this for http://www.microsoft.com/windows: field:(http OR www OR microsoft OR come OR windows) For things like that, you'll need to build your own queries, not use the parser. -Kurt -----Original Message----- From: Patrick Burrows [mailto:[EMAIL PROTECTED] Sent: Tuesday, July 31, 2007 1:45 PM To: [email protected] Subject: Re: test hmm.... this seems to have made it through. My previous posts kept getting bounced for being spam. I had been trying to ask a question on searching for URLs. One of the fields in my index is called link. It holds nothing but URLs. There may be more than one link field per document. When I search on the url, though (using field:fullurl syntax) it returns a hit on every field in the database. Is there special syntax for searching for a url? On 7/31/07, Patrick Burrows <mailto:[EMAIL PROTECTED]> <[EMAIL PROTECTED]> wrote: this is not spam, please stop bouncing it -- - P -- - P -- - P -- - P __________ NOD32 2220 (20070426) Information __________ This message was checked by NOD32 antivirus system. http://www.eset.com
