why would Explanation be *ALWAYS* null?

[ro...@earthlink.net]

R2.4 

There is much about Lucene that I do not understand, so it may be that there is some simple or obvious mistake I am making. I build an index, get hits (documents) back from it, with various non-zero scores. Now I call this code: 

	Explanation expl = _searcher.explain(rewrite, docIndex); // always null! 

And the return value is *ALWAYS* null. The javadocs on explain() do not even mention that possibility. When I debug through the call, I find the "explanation" in this code inside class MarkupContainsQuery (which is the code that gets called): 

		/**
		 * @see Weight#explain(IndexReader, int)
		 */
		public Explanation explain(IndexReader reader, int doc)
				throws IOException {
			if (reader == null) {
				throw new IOException("Empty reader in explain method");
			}
			// TODO SY - implement
>>>>>>>>>		return null;
		}

This seems, um, bad, esp. as I did nothing to *decide* the type of my query, but merely created it from the user-entered string like this: 

	Query query = _parser.parse(queryString);
	Query rewrite = query.rewrite(_reader);

And the field _parser was created like this: 

	Analyzer analyzer = new StandardAnalyzer();
	QueryParser parser = new QueryParser("contents", analyzer);
	_parser XMLFragmentParser(parser);

So -- what's the right thing to do? 

thanks,
Paul 




---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org

Re: why would Explanation be *ALWAYS* null?

[Chris Hostetter <ho...@fucit.org>]

: even mention that possibility. When I debug through the call, I find the 
: "explanation" in this code inside class MarkupContainsQuery (which is 
: the code that gets called):
	...
: 			// TODO SY - implement
: >>>>>>>>>		return null;
: 		}

that's definitely not good ... but (unless i'm a *LOT* more tired then 
i think) it's also not in the Lucene code base.

: This seems, um, bad, esp. as I did nothing to *decide* the type of my 
: query, but merely created it from the user-entered string like this:
: 
: 	Query query = _parser.parse(queryString);
: 	Query rewrite = query.rewrite(_reader);
: 
: And the field _parser was created like this: 
: 
: 	Analyzer analyzer = new StandardAnalyzer();
: 	QueryParser parser = new QueryParser("contents", analyzer);
: 	_parser XMLFragmentParser(parser);

I don't really understand that code ... it doesn't seem to be legal java. 
is that last line suppose to be an assignment to a _parser variable from 
a constructor to some class named XMLFragmentParser ?

where did you get XMLFragmentParser from?  I'm seeing mentions here that 
seem like they might be applicable, but i'm not fiding any code...
    http://riao.free.fr/papers/15.pdf

: So -- what's the right thing to do? 

I'd file a bug with whoever provided you the XMLFragmentParser and 
MarkupContainsQuery classes.


-Hoss


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org

Re: why would Explanation be *ALWAYS* null?

[Ian Lea <ia...@gmail.com>]

Have you tried it without the rewrite, marked as "expert" in the javadocs?

I'm not a lucene expert but I am an experienced user and in several
years of usage I don't think I've ever had cause to call
Query.rewrite().


--
Ian.

On Wed, Jan 21, 2009 at 7:44 PM,  <ro...@earthlink.net> wrote:
> R2.4
>
> There is much about Lucene that I do not understand, so it may be that there is some simple or obvious mistake I am making. I build an index, get hits (documents) back from it, with various non-zero scores. Now I call this code:
>
>        Explanation expl = _searcher.explain(rewrite, docIndex); // always null!
>
> And the return value is *ALWAYS* null. The javadocs on explain() do not even mention that possibility. When I debug through the call, I find the "explanation" in this code inside class MarkupContainsQuery (which is the code that gets called):
>
>                /**
>                 * @see Weight#explain(IndexReader, int)
>                 */
>                public Explanation explain(IndexReader reader, int doc)
>                                throws IOException {
>                        if (reader == null) {
>                                throw new IOException("Empty reader in explain method");
>                        }
>                        // TODO SY - implement
>>>>>>>>>>               return null;
>                }
>
> This seems, um, bad, esp. as I did nothing to *decide* the type of my query, but merely created it from the user-entered string like this:
>
>        Query query = _parser.parse(queryString);
>        Query rewrite = query.rewrite(_reader);
>
> And the field _parser was created like this:
>
>        Analyzer analyzer = new StandardAnalyzer();
>        QueryParser parser = new QueryParser("contents", analyzer);
>        _parser XMLFragmentParser(parser);
>
> So -- what's the right thing to do?
>
> thanks,
> Paul
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org