lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jan Høydahl / Cominvent <jan....@cominvent.com>
Subject Re: dismax and AND as the default operator
Date Fri, 18 Jun 2010 16:52:12 GMT
Standard DisMax does not fully support explicit AND/OR.
You can prove that by trying to say q=fuel+OR+cell and see that the score stays the same (given
mm=100%)
It appears that DisMax does SOME intelligent handling of AND/OR/NOT, because it adds the "+"
on the AND and a "-" on the NOT. But adding a "+" is redundant and does not change anything
as long as mm=100%. The NOT actually seems to work, but the OR does not have any effect due
to the "+" on the top-level ().

If you need boolean syntax support in DisMax, try the defType=edismax with patch SOLR-1553
or alternatively on branch_3x

--
Jan Høydahl, search solution architect
Cominvent AS - www.cominvent.com
Training in Europe - www.solrtraining.com

On 18. juni 2010, at 02.44, Erik Hatcher wrote:

> Hmmm, maybe I'm wrong and it does support AND.  Looking at the code I don't see why it
wouldn't, actually.  Though I believe I've seen it documented that it isn't supported (or
at least not advertised to support).  Ok, from the dismax wiki page it says: "This query handler
supports an extremely simplified subset of the Lucene QueryParser syntax. Quotes can be used
to group phrases, and +/- can be used to denote mandatory and optional clauses".  Only special
single characters are escaped.  So AND/OR must work.  Learn something new every day!
> 
> 	Erik
> 
> 
> 
> On Jun 17, 2010, at 8:28 PM, Tommy Chheng wrote:
> 
>> Thanks, Erik. that does work. I misunderstood the documentation, i thought "clause"
meant "field" rather than the terms in the query.
>> 
>> If dismax doesn't support the operator AND,  why would the query "solr/select?q=fuel+cell"
and "solr/select?q=fuel+AND+cell" get parsed differently(it adds the + for the AND query)
and have different result count?
>> 
>> @tommychheng
>> Programmer and UC Irvine Graduate Student
>> Find a great grad school based on research interests: http://gradschoolnow.com
>> 
>> 
>> On 6/17/10 5:17 PM, Erik Hatcher wrote:
>>> dismax does not support the operator AND.  It uses +/- only.
>>> 
>>> set mm=100% (not 1), as Hoss said, and try your query again.
>>> 
>>>   Erik
>>> 
>>> On Jun 17, 2010, at 8:08 PM, Tommy Chheng wrote:
>>> 
>>>> I don't think setting the mm helps.
>>>> I have mm to 1 which means the query terms should be in at least one field.
Both query strings satisfy this condition.
>>>> 
>>>> The query "solr/select?q=fuel+cell" is parsed as
>>>> 
>>>> "querystring":"fuel cell",
>>>> "parsedquery":"+((DisjunctionMaxQuery((text:fuel | organization_name_ws_lc:fuel^5.0))
DisjunctionMaxQuery((text:cell | organization_name_ws_lc:cell^5.0)))~1) ()",
>>>> "parsedquery_toString":"+(((text:fuel | organization_name_ws_lc:fuel^5.0)
(text:cell | organization_name_ws_lc:cell^5.0))~1) ()",
>>>> 
>>>> returns ~900 results
>>>> 
>>>> The query "solr/select?q=fuel+AND+cell" is parsed as
>>>> 
>>>> "querystring":"fuel AND cell",
>>>> "parsedquery":"+(+DisjunctionMaxQuery((text:fuel | organization_name_ws_lc:fuel^5.0))
+DisjunctionMaxQuery((text:cell | organization_name_ws_lc:cell^5.0))) ()",
>>>> "parsedquery_toString":"+(+(text:fuel | organization_name_ws_lc:fuel^5.0)
+(text:cell | organization_name_ws_lc:cell^5.0)) ()",
>>>> returns ~80 results
>>>> 
>>>> (this is the behavior i want for query "fuel cell" because it adds the extra
+). I want to do this without adding the AND for every query.
>>>> 
>>>> 
>>>> 
>>>> @tommychheng
>>>> Programmer and UC Irvine Graduate Student
>>>> Find a great grad school based on research interests: http://gradschoolnow.com
>>>> 
>>>> 
>>>> On 6/17/10 4:19 PM, Chris Hostetter wrote:
>>>>> :  I'm using the dismax request handler and want to set the default operator
to
>>>>> : AND.
>>>>> : Using the standard handler, i could just use the q.op or defaultOperator
in
>>>>> : the schema, but this doesn't work using the dismax request handler.
>>>>> :
>>>>> : For example, if I call "solr/select/?q=fuel+cell", I want solr to handle
it as
>>>>> : a "solr/select/?q=fuel+AND+cell"
>>>>> 
>>>>> Please consult the dismax docs...
>>>>> http://wiki.apache.org/solr/DisMaxRequestHandler#mm_.28Minimum_.27Should.27_Match.29
>>>>> 
>>>>> dismax uses the "mm" param to decide how clauses that don't have an
>>>>> explicit operator will be dealt with -- the default is to require 100%
of
>>>>> the terms, so if you aren't seeing that behavior then you have a
>>>>> solrconfig.xml that that sets the default mm value to something else.
>>>>> 
>>>>> Starting with Solr 4.0 (and mybe 3.1 if it's backported) the default
mm
>>>>> will be based on the value of q.op (see SOLR-1889 for more details)
>>>>> 
>>>>> 
>>>>> -Hoss
>>>>> 
>>> 
> 


Mime
View raw message