You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Robin Wei <th...@163.com> on 2013/09/16 09:08:36 UTC
2 question about solr and lucene
Hi, guys:
I met two questions about solr and lucene, wish people to help out.
use payload query but can NOT with numerical field type. for example:
I implemented my own requesthandler, refer to http://hnagtech.wordpress.com/2013/04/19/using-payloads-with-solr-4-x/
I query in solr: sinaTag:operate
solr response:
"numFound": 2,
"start": 0, "maxScore": 99, "docs": [ { "id": "1628209010",
"followersCount": 752,
"sinaTag": "operate|99 C2C|98 B2C|97 OnlineShopping|96 E-commercial|94",
"score": 99
},
{ "id": "1900546410",
"followersCount": 1002,
"sinaTag": "Startup|99 Benz|98 PublicRelation|97 operate|96 Activity|95 Media|94 AD|93 Vehicle|92 ",
"score": 96
}
This work well.
But query with combined with other numberical condition, such as:
sinaTag:operate and followersCount:[752 TO 752]
{ "responseHeader": { "status": 0, "QTime": 40 }, "response": { "numFound": 0, "start": 0, "maxScore": 0, "docs": [] }
}
According these dataset, the first record should be responsed rather than NOT FOUND.
I not know why.
2. About string field fuzzy match filtering, how to get the score? what the formula is?
When I used two or several string fuzzy match, probable AND or OR, how to get the score? what the formula is?
Might I implement myself score formula class which interface or abstract class to extend ?
Thanks in advance.
Re: 2 question about solr and lucene
Posted by Erick Erickson <er...@gmail.com>.
Are you saying you're trying to put payloads on the numeric data?
If that's the case I don't know how that works. But a couple of things:
sinaTag:operate and followersCount:[752 TO 752]
is incorrect, you must capitalize the and as
sinaTag:operate AND followersCount:[752 TO 752]
Your syntax _should_ work since [] is inclusive, but
I'd just try with 751 TO 753] once to be sure.
Attach &debug=all to your query and you'll see exactly how the
query is parsed. You'll also see exactly how the scores are
calculated in a long, complex bit of output. I _think_ that
fuzzy and wildcards do a "constant score query", so don't be
surprised if the calculations show you that the fuzzy matches don't
change the score.
Best,
Erick
On Mon, Sep 16, 2013 at 3:08 AM, Robin Wei <th...@163.com> wrote:
> Hi, guys:
> I met two questions about solr and lucene, wish people to help out.
>
> use payload query but can NOT with numerical field type. for example:
> I implemented my own requesthandler, refer to
> http://hnagtech.wordpress.com/2013/04/19/using-payloads-with-solr-4-x/
> I query in solr: sinaTag:operate
> solr response:
>
> "numFound": 2,
> "start": 0, "maxScore": 99, "docs": [ { "id":
> "1628209010",
> "followersCount": 752,
> "sinaTag": "operate|99 C2C|98 B2C|97 OnlineShopping|96
> E-commercial|94",
> "score": 99
>
> },
>
> { "id": "1900546410",
> "followersCount": 1002,
> "sinaTag": "Startup|99 Benz|98 PublicRelation|97 operate|96
> Activity|95 Media|94 AD|93 Vehicle|92 ",
> "score": 96
>
> }
>
> This work well.
> But query with combined with other numberical condition, such as:
> sinaTag:operate and followersCount:[752 TO 752]
> { "responseHeader": { "status": 0, "QTime":
> 40 }, "response": { "numFound": 0, "start": 0,
> "maxScore": 0, "docs": [] }
> }
> According these dataset, the first record should be responsed rather
> than NOT FOUND.
> I not know why.
>
>
> 2. About string field fuzzy match filtering, how to get the score? what
> the formula is?
> When I used two or several string fuzzy match, probable AND or OR,
> how to get the score? what the formula is?
> Might I implement myself score formula class which interface or
> abstract class to extend ?
>
>
>
>
>
> Thanks in advance.
>
>
>
>
>