sokolov at ifactory
May 6, 2012, 5:22 PM
Post #6 of 6
(166 views)
Permalink
|
Hmm - I looked at Spans more carefully, and it appears as if your idea
about a "fake" Query (some kind of SpanAllQuery would be called for)
would work well, and would probably be much simpler to implement. It
wouldn't preclude the kind of optimization I was talking about either,
but I don't know if it would be worth the trouble.
It turns out in my very specific case I have a term that appears in
every document in a particular field, so I am just using a search for
that at the moment.
-Mike
On 5/6/2012 8:04 PM, Mike Sokolov wrote:
> I think what I have in mind would be purely an artifact of the parser;
> a term that would always be optimized away, either vanishing or
> gobbling up the whole query. So if you had n(A,*), you would just
> get "A". If you had and(A, not(*)) (is that the surround syntax for
> not?) you would get nothing, if you had * you would get all the
> documents. Maybe this could be done without having to actually
> generate a query internally, but could happen during parsing. It's
> kind of a weird case, but I am trying to translate from one query
> language to another, and it would be convenient to have this as an
> option.
>
> -Mike
>
> On 5/6/2012 7:28 PM, Robert Muir wrote:
>> Hi Mike: wheres for the normal queryparser this Query doesn't consult
>> the positions file and is trivial, how would such a query be
>> implemented for the surround parser? As a single span that matches all
>> positions for the whole document? Maybe it could be a "fake span" for
>> each document of 0 ... Integer.MAX_VALUE?
>>
>> I think it would be nice to have as long as its not going to be too
>> inefficient...
>>
>> On Sun, May 6, 2012 at 5:26 PM, Mike Sokolov<sokolov [at] ifactory>
>> wrote:
>>> does anybody know how to express a MatchAllDocsQuery in surround query
>>> parser language? I've tried
>>>
>>> *
>>>
>>> and()
>>>
>>> but those don't parse. I looked at the grammar and I don't think
>>> there is a
>>> way. Please let us all know if you know otherwise!
>>>
>>> Thanks
>>>
>>> -Mike Sokolov
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: java-user-unsubscribe [at] lucene
>>> For additional commands, e-mail: java-user-help [at] lucene
>>>
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe [at] lucene
> For additional commands, e-mail: java-user-help [at] lucene
>
---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe [at] lucene
For additional commands, e-mail: java-user-help [at] lucene
|
