Compare scores from multiple indices

2014-04-16 Thread Sven Teichmann
Hello everyone, because of the large amount of data we have to index, we do not create one large index but instead multiple small indices (we call them volumes) which will then be put on multiple servers (our search cluster). If I understand Lucene's scoring correctly, I am not able to compar

Re: is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Jose Carlos Canova
Usually when i search for something, i know exactly what i am expecting to find, the user that don't know what will find isn't searching for nothing, since random search will provide random walks over a certain result path, I believe that search generally speaking is the expectation of some agreeme

Nonprofit Staffing Discount

2014-04-16 Thread nonprofits
To support your mission and staffing efforts, Nonprofit People & Systems has arranged a special coupon code which will allow you to post UNLIMITED jobs at www.nonprofitJOBMARKET.org all year long for a one-time fee of $50 (code: PS2014 to be used at checkout)! You may share your coupon code

BooleanScorer - Maximum Prohibited Scorers?

2014-04-16 Thread David Stimpert
Hello, I have found useful functionality in BooleanQuery which allows me to specify a minimum number of matching optional terms (i.e. setMinimumNumberShouldMatch). I do not, however, see similar functionality available for setting the maximum number of MUST_NOTs (i.e. setMaximumNumberMustNotMatch)

Re: is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Jack Krupansky
By definition "OR" is the "disjunction" operator, and "AND" is the "conjunction" operator. Yes, the default operator is "OR" (the disjunction operator.) The original question could have been phrased as "Why is recall more important than precision?" The answer to that is rooted in the fact th

Re: is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Chris Hostetter
: I recently wondered, : why lucene's default conjunction operator is "OR". : Is there a historical reason for that? The only 'default' is in the query parser -- if you construct the BooleanQueyr objects programatically you must always be explicit about the Occur property of each Clause. In t

Re: is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Jose Carlos Canova
In fact you have both, the documents at see looking at first time is first the results with all words (AND) then the ORed results, which makes perfect sense. Google sometimes marks on the result which word was not found with a "strike through". But it is not so powerful as logical operators on qu

Re: is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Herb Roitblat
Actually, Google uses OR. The scoring algorithm favors documents that match on more of the ORed terms. On 4/16/2014 8:17 AM, Min-Uk Kim wrote: Hello everyone, I recently wondered, why lucene's default conjunction operator is "OR". Is there a historical reason for that? By the way, Google an

is there a historical reason why default conjunction operator is "OR"?

2014-04-16 Thread Min-Uk Kim
Hello everyone, I recently wondered, why lucene's default conjunction operator is "OR". Is there a historical reason for that? By the way, Google and other search engines seem to use "AND". Please show me the light. M

Re: IndexReplication Client and IndexWriter

2014-04-16 Thread Christoph Kaser
Hi Shai, no problem, thank you for letting me know! Have a nice vacation! Christoph Am 16.04.2014 08:49, schrieb Shai Erera: Hi Christoph, Apologize for the delayed response, I'm on a holiday vacation. I will take a look at your issues as soon as I can. Shai On Fri, Apr 11, 2014 at 12:02

Re: Indexing Huge tree structure represented in a Text file

2014-04-16 Thread kumagirish
hello Arjen van der Meijden if its not too much of a trouble can you point me to any sites with example implementation on Neo4j for problem similar to mine i want to check if neo4j resolve all my problems as this is new technology i need to do a lot of research and feew examples will be a good st

Re: AnalyzingInfixSuggester.build() corrupts index for plain searching?

2014-04-16 Thread Michael McCandless
You should not use the same indexDir for your "normal" Lucene index (for searching), and for the AnalyzingInfixSuggester. The two are completely different, and AnalyzingInfixSuggester happily overwrites the entire index you passed it when you call its build method. I'll fix the javadocs to try to