Impossible Boost Query?

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Impossible Boost Query?

Blargy
I was wondering if this is even possible. I'll try to explain what I'm trying to do to the best of my ability.

Ok, so our site has a bunch of products that are sold by any number of sellers. Currently when I search for some product I get back all products matching that search term but the problem is there may be multiple products sold by the same seller that are all closely related, therefore their scores are related. So basically the search ends up with results that are all closely clumped together by the same seller but I would much rather prefer to distribute these results across sellers (given each seller a fair shot to sell their goods).

Is there any way to add some boost query for example that will start weighing products lower when their seller has already been listed a few times. For example, right now I have

Product foo by Seller A
Product foo by Seller A
Product foo by Seller A
Product foo by Seller B
Product foo by Seller B
Product foo by Seller B
Product foo by Seller C
Product foo by Seller C
Product foo by Seller C

where each result is very close in score. I would like something like this

Product foo by Seller A
Product foo by Seller B
Product foo by Seller C
Product foo by Seller A
Product foo by Seller B
Product foo by Seller C
....

basically distributing the results over the sellers. Is something like this possible? I don't care if the solution involves a boost query or not. I just want some way to distribute closely related documents.

Thanks!!!
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Otis Gospodnetic-2
Would Field Collapsing from SOLR-236 do the job for you?

Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/



----- Original Message ----
> From: blargy <[hidden email]>
> To: [hidden email]
> Sent: Tue, March 23, 2010 2:39:48 PM
> Subject: Impossible Boost Query?
>
>
I was wondering if this is even possible. I'll try to explain what I'm
> trying
to do to the best of my ability.

Ok, so our site has a bunch
> of products that are sold by any number of
sellers. Currently when I search
> for some product I get back all products
matching that search term but the
> problem is there may be multiple products
sold by the same seller that are
> all closely related, therefore their scores
are related. So basically the
> search ends up with results that are all
closely clumped together by the same
> seller but I would much rather prefer
to distribute these results across
> sellers (given each seller a fair shot to
sell their goods).

Is there
> any way to add some boost query for example that will start
weighing products
> lower when their seller has already been listed a few
times. For example,
> right now I have

Product foo by Seller A
Product foo by Seller
> A
Product foo by Seller A
Product foo by Seller B
Product foo by Seller
> B
Product foo by Seller B
Product foo by Seller C
Product foo by Seller
> C
Product foo by Seller C

where each result is very close in score. I
> would like something like this

Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
....

basically distributing the
> results over the sellers. Is something like this
possible? I don't care if
> the solution involves a boost query or not. I just
want some way to
> distribute closely related documents.

Thanks!!!
--
View this
> message in context:
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
Sent
> from the Solr - User mailing list archive at Nabble.com.
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Blargy
Possibly. How can I install this as a contrib or do I need to actually perform the patch?

Otis Gospodnetic wrote
Would Field Collapsing from SOLR-236 do the job for you?

Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/



----- Original Message ----
> From: blargy <zmanods@hotmail.com>
> To: solr-user@lucene.apache.org
> Sent: Tue, March 23, 2010 2:39:48 PM
> Subject: Impossible Boost Query?
>
>
I was wondering if this is even possible. I'll try to explain what I'm
> trying
to do to the best of my ability.

Ok, so our site has a bunch
> of products that are sold by any number of
sellers. Currently when I search
> for some product I get back all products
matching that search term but the
> problem is there may be multiple products
sold by the same seller that are
> all closely related, therefore their scores
are related. So basically the
> search ends up with results that are all
closely clumped together by the same
> seller but I would much rather prefer
to distribute these results across
> sellers (given each seller a fair shot to
sell their goods).

Is there
> any way to add some boost query for example that will start
weighing products
> lower when their seller has already been listed a few
times. For example,
> right now I have

Product foo by Seller A
Product foo by Seller
> A
Product foo by Seller A
Product foo by Seller B
Product foo by Seller
> B
Product foo by Seller B
Product foo by Seller C
Product foo by Seller
> C
Product foo by Seller C

where each result is very close in score. I
> would like something like this

Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
....

basically distributing the
> results over the sellers. Is something like this
possible? I don't care if
> the solution involves a boost query or not. I just
want some way to
> distribute closely related documents.

Thanks!!!
--
View this
> message in context:
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
Sent
> from the Solr - User mailing list archive at Nabble.com.
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Blargy
Maybe a better question is... how can I install this and will it work with 1.4?

Thanks

blargy wrote
Possibly. How can I install this as a contrib or do I need to actually perform the patch?

Otis Gospodnetic wrote
Would Field Collapsing from SOLR-236 do the job for you?

Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/



----- Original Message ----
> From: blargy <zmanods@hotmail.com>
> To: solr-user@lucene.apache.org
> Sent: Tue, March 23, 2010 2:39:48 PM
> Subject: Impossible Boost Query?
>
>
I was wondering if this is even possible. I'll try to explain what I'm
> trying
to do to the best of my ability.

Ok, so our site has a bunch
> of products that are sold by any number of
sellers. Currently when I search
> for some product I get back all products
matching that search term but the
> problem is there may be multiple products
sold by the same seller that are
> all closely related, therefore their scores
are related. So basically the
> search ends up with results that are all
closely clumped together by the same
> seller but I would much rather prefer
to distribute these results across
> sellers (given each seller a fair shot to
sell their goods).

Is there
> any way to add some boost query for example that will start
weighing products
> lower when their seller has already been listed a few
times. For example,
> right now I have

Product foo by Seller A
Product foo by Seller
> A
Product foo by Seller A
Product foo by Seller B
Product foo by Seller
> B
Product foo by Seller B
Product foo by Seller C
Product foo by Seller
> C
Product foo by Seller C

where each result is very close in score. I
> would like something like this

Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
Product foo by Seller A
Product foo by
> Seller B
Product foo by Seller C
....

basically distributing the
> results over the sellers. Is something like this
possible? I don't care if
> the solution involves a boost query or not. I just
want some way to
> distribute closely related documents.

Thanks!!!
--
View this
> message in context:
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
Sent
> from the Solr - User mailing list archive at Nabble.com.
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Otis Gospodnetic-2
You'd likely want to get the latest patch and trunk and try applying.

Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/



----- Original Message ----
> From: blargy <[hidden email]>
> To: [hidden email]
> Sent: Tue, March 23, 2010 6:10:22 PM
> Subject: Re: Impossible Boost Query?
>
>
Maybe a better question is... how can I install this and will it work
> with
1.4?

Thanks


blargy wrote:

>
> Possibly.
> How can I install this as a contrib or do I need to actually
> perform the
> patch?
>
>
> Otis Gospodnetic wrote:
>>
>
>> Would Field Collapsing from SOLR-236 do the job for
> you?
>>
>> Otis
>> ----
>> Sematext ::
> href="http://sematext.com/" target=_blank >http://sematext.com/ :: Solr -
> Lucene - Nutch
>> Hadoop ecosystem search ::
> href="http://search-hadoop.com/" target=_blank
> >http://search-hadoop.com/
>>
>>
>>
>
>> ----- Original Message ----
>>> From: blargy <
> ymailto="mailto:[hidden email]"
> href="mailto:[hidden email]">[hidden email]>
>>>
> To:
> href="mailto:[hidden email]">[hidden email]
>>>
> Sent: Tue, March 23, 2010 2:39:48 PM
>>> Subject: Impossible Boost
> Query?
>>>
>>>
>> I was wondering if this is
> even possible. I'll try to explain what I'm
>>> trying
>>
> to do to the best of my ability.
>>
>> Ok, so our site has a
> bunch
>>> of products that are sold by any number of
>>
> sellers. Currently when I search
>>> for some product I get back
> all products
>> matching that search term but the
>>>
> problem is there may be multiple products
>> sold by the same seller
> that are
>>> all closely related, therefore their
> scores
>> are related. So basically the
>>> search ends up
> with results that are all
>> closely clumped together by the same
>
>>> seller but I would much rather prefer
>> to distribute
> these results across
>>> sellers (given each seller a fair shot
> to
>> sell their goods).
>>
>> Is there
>
>>> any way to add some boost query for example that will
> start
>> weighing products
>>> lower when their seller has
> already been listed a few
>> times. For example,
>>> right
> now I have
>>
>> Product foo by Seller A
>> Product
> foo by Seller
>>> A
>> Product foo by Seller A
>>
> Product foo by Seller B
>> Product foo by Seller
>>>
> B
>> Product foo by Seller B
>> Product foo by Seller
> C
>> Product foo by Seller
>>> C
>> Product foo
> by Seller C
>>
>> where each result is very close in score. I
>
>>> would like something like this
>>
>> Product
> foo by Seller A
>> Product foo by
>>> Seller B
>>
> Product foo by Seller C
>> Product foo by Seller A
>> Product
> foo by
>>> Seller B
>> Product foo by Seller C
>>
> ....
>>
>> basically distributing the
>>>
> results over the sellers. Is something like this
>> possible? I don't
> care if
>>> the solution involves a boost query or not. I
> just
>> want some way to
>>> distribute closely related
> documents.
>>
>> Thanks!!!
>> --
>> View
> this
>>> message in context:
>>> href="
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>
>>> target=_blank
>>> >
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>
> Sent
>>> from the Solr - User mailing list archive at
> Nabble.com.
>>
>>
>
>

--
View this
> message in context:
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
Sent
> from the Solr - User mailing list archive at Nabble.com.
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Blargy
Thanks but Im not quite show on how to apply the patch. I just use the packaged solr-1.4.0.war in my deployment (no compiling, etc). Is there a way I can patch the war file?

Any instructions would be greatly appreciated. Thanks

Otis Gospodnetic wrote
You'd likely want to get the latest patch and trunk and try applying.

Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/



----- Original Message ----
> From: blargy <zmanods@hotmail.com>
> To: solr-user@lucene.apache.org
> Sent: Tue, March 23, 2010 6:10:22 PM
> Subject: Re: Impossible Boost Query?
>
>
Maybe a better question is... how can I install this and will it work
> with
1.4?

Thanks


blargy wrote:
>
> Possibly.
> How can I install this as a contrib or do I need to actually
> perform the
> patch?
>
>
> Otis Gospodnetic wrote:
>>
>
>> Would Field Collapsing from SOLR-236 do the job for
> you?
>>
>> Otis
>> ----
>> Sematext ::
> href="http://sematext.com/" target=_blank >http://sematext.com/ :: Solr -
> Lucene - Nutch
>> Hadoop ecosystem search ::
> href="http://search-hadoop.com/" target=_blank
> >http://search-hadoop.com/
>>
>>
>>
>
>> ----- Original Message ----
>>> From: blargy <
> ymailto="mailto:zmanods@hotmail.com"
> href="mailto:zmanods@hotmail.com">zmanods@hotmail.com>
>>>
> To:
> href="mailto:solr-user@lucene.apache.org">solr-user@lucene.apache.org
>>>
> Sent: Tue, March 23, 2010 2:39:48 PM
>>> Subject: Impossible Boost
> Query?
>>>
>>>
>> I was wondering if this is
> even possible. I'll try to explain what I'm
>>> trying
>>
> to do to the best of my ability.
>>
>> Ok, so our site has a
> bunch
>>> of products that are sold by any number of
>>
> sellers. Currently when I search
>>> for some product I get back
> all products
>> matching that search term but the
>>>
> problem is there may be multiple products
>> sold by the same seller
> that are
>>> all closely related, therefore their
> scores
>> are related. So basically the
>>> search ends up
> with results that are all
>> closely clumped together by the same
>
>>> seller but I would much rather prefer
>> to distribute
> these results across
>>> sellers (given each seller a fair shot
> to
>> sell their goods).
>>
>> Is there
>
>>> any way to add some boost query for example that will
> start
>> weighing products
>>> lower when their seller has
> already been listed a few
>> times. For example,
>>> right
> now I have
>>
>> Product foo by Seller A
>> Product
> foo by Seller
>>> A
>> Product foo by Seller A
>>
> Product foo by Seller B
>> Product foo by Seller
>>>
> B
>> Product foo by Seller B
>> Product foo by Seller
> C
>> Product foo by Seller
>>> C
>> Product foo
> by Seller C
>>
>> where each result is very close in score. I
>
>>> would like something like this
>>
>> Product
> foo by Seller A
>> Product foo by
>>> Seller B
>>
> Product foo by Seller C
>> Product foo by Seller A
>> Product
> foo by
>>> Seller B
>> Product foo by Seller C
>>
> ....
>>
>> basically distributing the
>>>
> results over the sellers. Is something like this
>> possible? I don't
> care if
>>> the solution involves a boost query or not. I
> just
>> want some way to
>>> distribute closely related
> documents.
>>
>> Thanks!!!
>> --
>> View
> this
>>> message in context:
>>> href="
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>
>>> target=_blank
>>> >
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>
> Sent
>>> from the Solr - User mailing list archive at
> Nabble.com.
>>
>>
>
>

--
View this
> message in context:
> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
> target=_blank
> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
Sent
> from the Solr - User mailing list archive at Nabble.com.
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Lance Norskog-2
At this point (and for almost 3 years :) field collapsing is a source
patch. You have to check out the Solr trunk from the Apache subversion
server, apply the patch with the 'patch' command, and build the new
Solr with 'ant'.

On Tue, Mar 23, 2010 at 4:13 PM, blargy <[hidden email]> wrote:

>
> Thanks but Im not quite show on how to apply the patch. I just use the
> packaged solr-1.4.0.war in my deployment (no compiling, etc). Is there a way
> I can patch the war file?
>
> Any instructions would be greatly appreciated. Thanks
>
>
> Otis Gospodnetic wrote:
>>
>> You'd likely want to get the latest patch and trunk and try applying.
>>
>> Otis
>> ----
>> Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
>> Hadoop ecosystem search :: http://search-hadoop.com/
>>
>>
>>
>> ----- Original Message ----
>>> From: blargy <[hidden email]>
>>> To: [hidden email]
>>> Sent: Tue, March 23, 2010 6:10:22 PM
>>> Subject: Re: Impossible Boost Query?
>>>
>>>
>> Maybe a better question is... how can I install this and will it work
>>> with
>> 1.4?
>>
>> Thanks
>>
>>
>> blargy wrote:
>>>
>>> Possibly.
>>> How can I install this as a contrib or do I need to actually
>>> perform the
>>> patch?
>>>
>>>
>>> Otis Gospodnetic wrote:
>>>>
>>>
>>>> Would Field Collapsing from SOLR-236 do the job for
>>> you?
>>>>
>>>> Otis
>>>> ----
>>>> Sematext ::
>>> href="http://sematext.com/" target=_blank >http://sematext.com/ :: Solr -
>>> Lucene - Nutch
>>>> Hadoop ecosystem search ::
>>> href="http://search-hadoop.com/" target=_blank
>>> >http://search-hadoop.com/
>>>>
>>>>
>>>>
>>>
>>>> ----- Original Message ----
>>>>> From: blargy <
>>> ymailto="mailto:[hidden email]"
>>> href="mailto:[hidden email]">[hidden email]>
>>>>>
>>> To:
>>> href="mailto:[hidden email]">[hidden email]
>>>>>
>>> Sent: Tue, March 23, 2010 2:39:48 PM
>>>>> Subject: Impossible Boost
>>> Query?
>>>>>
>>>>>
>>>> I was wondering if this is
>>> even possible. I'll try to explain what I'm
>>>>> trying
>>>>
>>> to do to the best of my ability.
>>>>
>>>> Ok, so our site has a
>>> bunch
>>>>> of products that are sold by any number of
>>>>
>>> sellers. Currently when I search
>>>>> for some product I get back
>>> all products
>>>> matching that search term but the
>>>>>
>>> problem is there may be multiple products
>>>> sold by the same seller
>>> that are
>>>>> all closely related, therefore their
>>> scores
>>>> are related. So basically the
>>>>> search ends up
>>> with results that are all
>>>> closely clumped together by the same
>>>
>>>>> seller but I would much rather prefer
>>>> to distribute
>>> these results across
>>>>> sellers (given each seller a fair shot
>>> to
>>>> sell their goods).
>>>>
>>>> Is there
>>>
>>>>> any way to add some boost query for example that will
>>> start
>>>> weighing products
>>>>> lower when their seller has
>>> already been listed a few
>>>> times. For example,
>>>>> right
>>> now I have
>>>>
>>>> Product foo by Seller A
>>>> Product
>>> foo by Seller
>>>>> A
>>>> Product foo by Seller A
>>>>
>>> Product foo by Seller B
>>>> Product foo by Seller
>>>>>
>>> B
>>>> Product foo by Seller B
>>>> Product foo by Seller
>>> C
>>>> Product foo by Seller
>>>>> C
>>>> Product foo
>>> by Seller C
>>>>
>>>> where each result is very close in score. I
>>>
>>>>> would like something like this
>>>>
>>>> Product
>>> foo by Seller A
>>>> Product foo by
>>>>> Seller B
>>>>
>>> Product foo by Seller C
>>>> Product foo by Seller A
>>>> Product
>>> foo by
>>>>> Seller B
>>>> Product foo by Seller C
>>>>
>>> ....
>>>>
>>>> basically distributing the
>>>>>
>>> results over the sellers. Is something like this
>>>> possible? I don't
>>> care if
>>>>> the solution involves a boost query or not. I
>>> just
>>>> want some way to
>>>>> distribute closely related
>>> documents.
>>>>
>>>> Thanks!!!
>>>> --
>>>> View
>>> this
>>>>> message in context:
>>>>> href="
>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>> target=_blank
>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>
>>>>> target=_blank
>>>>> >
>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>> target=_blank
>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>>>
>>> Sent
>>>>> from the Solr - User mailing list archive at
>>> Nabble.com.
>>>>
>>>>
>>>
>>>
>>
>> --
>> View this
>>> message in context:
>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
>>> target=_blank
>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
>> Sent
>>> from the Solr - User mailing list archive at Nabble.com.
>>
>>
>
> --
> View this message in context: http://old.nabble.com/Impossible-Boost-Query--tp28005354p28008495.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
>



--
Lance Norskog
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Lance Norskog-2
Also, there is a 'random' type which generates random numbers. This
might help you also.

On Tue, Mar 23, 2010 at 7:18 PM, Lance Norskog <[hidden email]> wrote:

> At this point (and for almost 3 years :) field collapsing is a source
> patch. You have to check out the Solr trunk from the Apache subversion
> server, apply the patch with the 'patch' command, and build the new
> Solr with 'ant'.
>
> On Tue, Mar 23, 2010 at 4:13 PM, blargy <[hidden email]> wrote:
>>
>> Thanks but Im not quite show on how to apply the patch. I just use the
>> packaged solr-1.4.0.war in my deployment (no compiling, etc). Is there a way
>> I can patch the war file?
>>
>> Any instructions would be greatly appreciated. Thanks
>>
>>
>> Otis Gospodnetic wrote:
>>>
>>> You'd likely want to get the latest patch and trunk and try applying.
>>>
>>> Otis
>>> ----
>>> Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
>>> Hadoop ecosystem search :: http://search-hadoop.com/
>>>
>>>
>>>
>>> ----- Original Message ----
>>>> From: blargy <[hidden email]>
>>>> To: [hidden email]
>>>> Sent: Tue, March 23, 2010 6:10:22 PM
>>>> Subject: Re: Impossible Boost Query?
>>>>
>>>>
>>> Maybe a better question is... how can I install this and will it work
>>>> with
>>> 1.4?
>>>
>>> Thanks
>>>
>>>
>>> blargy wrote:
>>>>
>>>> Possibly.
>>>> How can I install this as a contrib or do I need to actually
>>>> perform the
>>>> patch?
>>>>
>>>>
>>>> Otis Gospodnetic wrote:
>>>>>
>>>>
>>>>> Would Field Collapsing from SOLR-236 do the job for
>>>> you?
>>>>>
>>>>> Otis
>>>>> ----
>>>>> Sematext ::
>>>> href="http://sematext.com/" target=_blank >http://sematext.com/ :: Solr -
>>>> Lucene - Nutch
>>>>> Hadoop ecosystem search ::
>>>> href="http://search-hadoop.com/" target=_blank
>>>> >http://search-hadoop.com/
>>>>>
>>>>>
>>>>>
>>>>
>>>>> ----- Original Message ----
>>>>>> From: blargy <
>>>> ymailto="mailto:[hidden email]"
>>>> href="mailto:[hidden email]">[hidden email]>
>>>>>>
>>>> To:
>>>> href="mailto:[hidden email]">[hidden email]
>>>>>>
>>>> Sent: Tue, March 23, 2010 2:39:48 PM
>>>>>> Subject: Impossible Boost
>>>> Query?
>>>>>>
>>>>>>
>>>>> I was wondering if this is
>>>> even possible. I'll try to explain what I'm
>>>>>> trying
>>>>>
>>>> to do to the best of my ability.
>>>>>
>>>>> Ok, so our site has a
>>>> bunch
>>>>>> of products that are sold by any number of
>>>>>
>>>> sellers. Currently when I search
>>>>>> for some product I get back
>>>> all products
>>>>> matching that search term but the
>>>>>>
>>>> problem is there may be multiple products
>>>>> sold by the same seller
>>>> that are
>>>>>> all closely related, therefore their
>>>> scores
>>>>> are related. So basically the
>>>>>> search ends up
>>>> with results that are all
>>>>> closely clumped together by the same
>>>>
>>>>>> seller but I would much rather prefer
>>>>> to distribute
>>>> these results across
>>>>>> sellers (given each seller a fair shot
>>>> to
>>>>> sell their goods).
>>>>>
>>>>> Is there
>>>>
>>>>>> any way to add some boost query for example that will
>>>> start
>>>>> weighing products
>>>>>> lower when their seller has
>>>> already been listed a few
>>>>> times. For example,
>>>>>> right
>>>> now I have
>>>>>
>>>>> Product foo by Seller A
>>>>> Product
>>>> foo by Seller
>>>>>> A
>>>>> Product foo by Seller A
>>>>>
>>>> Product foo by Seller B
>>>>> Product foo by Seller
>>>>>>
>>>> B
>>>>> Product foo by Seller B
>>>>> Product foo by Seller
>>>> C
>>>>> Product foo by Seller
>>>>>> C
>>>>> Product foo
>>>> by Seller C
>>>>>
>>>>> where each result is very close in score. I
>>>>
>>>>>> would like something like this
>>>>>
>>>>> Product
>>>> foo by Seller A
>>>>> Product foo by
>>>>>> Seller B
>>>>>
>>>> Product foo by Seller C
>>>>> Product foo by Seller A
>>>>> Product
>>>> foo by
>>>>>> Seller B
>>>>> Product foo by Seller C
>>>>>
>>>> ....
>>>>>
>>>>> basically distributing the
>>>>>>
>>>> results over the sellers. Is something like this
>>>>> possible? I don't
>>>> care if
>>>>>> the solution involves a boost query or not. I
>>>> just
>>>>> want some way to
>>>>>> distribute closely related
>>>> documents.
>>>>>
>>>>> Thanks!!!
>>>>> --
>>>>> View
>>>> this
>>>>>> message in context:
>>>>>> href="
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>>
>>>>>> target=_blank
>>>>>> >
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>>>>
>>>> Sent
>>>>>> from the Solr - User mailing list archive at
>>>> Nabble.com.
>>>>>
>>>>>
>>>>
>>>>
>>>
>>> --
>>> View this
>>>> message in context:
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
>>> Sent
>>>> from the Solr - User mailing list archive at Nabble.com.
>>>
>>>
>>
>> --
>> View this message in context: http://old.nabble.com/Impossible-Boost-Query--tp28005354p28008495.html
>> Sent from the Solr - User mailing list archive at Nabble.com.
>>
>>
>
>
>
> --
> Lance Norskog
> [hidden email]
>



--
Lance Norskog
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Blargy
This sound a little closer to what I want but I don't want fully randomized results.

How exactly does this field work? Is it more than just a simple random sort (order by rand())? What would be nice is if I could randomize documents within a certain score percentage of each other. Is this available?

Thanks


Lance Norskog-2 wrote
Also, there is a 'random' type which generates random numbers. This
might help you also.

On Tue, Mar 23, 2010 at 7:18 PM, Lance Norskog <goksron@gmail.com> wrote:
> At this point (and for almost 3 years :) field collapsing is a source
> patch. You have to check out the Solr trunk from the Apache subversion
> server, apply the patch with the 'patch' command, and build the new
> Solr with 'ant'.
>
> On Tue, Mar 23, 2010 at 4:13 PM, blargy <zmanods@hotmail.com> wrote:
>>
>> Thanks but Im not quite show on how to apply the patch. I just use the
>> packaged solr-1.4.0.war in my deployment (no compiling, etc). Is there a way
>> I can patch the war file?
>>
>> Any instructions would be greatly appreciated. Thanks
>>
>>
>> Otis Gospodnetic wrote:
>>>
>>> You'd likely want to get the latest patch and trunk and try applying.
>>>
>>> Otis
>>> ----
>>> Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
>>> Hadoop ecosystem search :: http://search-hadoop.com/
>>>
>>>
>>>
>>> ----- Original Message ----
>>>> From: blargy <zmanods@hotmail.com>
>>>> To: solr-user@lucene.apache.org
>>>> Sent: Tue, March 23, 2010 6:10:22 PM
>>>> Subject: Re: Impossible Boost Query?
>>>>
>>>>
>>> Maybe a better question is... how can I install this and will it work
>>>> with
>>> 1.4?
>>>
>>> Thanks
>>>
>>>
>>> blargy wrote:
>>>>
>>>> Possibly.
>>>> How can I install this as a contrib or do I need to actually
>>>> perform the
>>>> patch?
>>>>
>>>>
>>>> Otis Gospodnetic wrote:
>>>>>
>>>>
>>>>> Would Field Collapsing from SOLR-236 do the job for
>>>> you?
>>>>>
>>>>> Otis
>>>>> ----
>>>>> Sematext ::
>>>> href="http://sematext.com/" target=_blank >http://sematext.com/ :: Solr -
>>>> Lucene - Nutch
>>>>> Hadoop ecosystem search ::
>>>> href="http://search-hadoop.com/" target=_blank
>>>> >http://search-hadoop.com/
>>>>>
>>>>>
>>>>>
>>>>
>>>>> ----- Original Message ----
>>>>>> From: blargy <
>>>> ymailto="mailto:zmanods@hotmail.com"
>>>> href="mailto:zmanods@hotmail.com">zmanods@hotmail.com>
>>>>>>
>>>> To:
>>>> href="mailto:solr-user@lucene.apache.org">solr-user@lucene.apache.org
>>>>>>
>>>> Sent: Tue, March 23, 2010 2:39:48 PM
>>>>>> Subject: Impossible Boost
>>>> Query?
>>>>>>
>>>>>>
>>>>> I was wondering if this is
>>>> even possible. I'll try to explain what I'm
>>>>>> trying
>>>>>
>>>> to do to the best of my ability.
>>>>>
>>>>> Ok, so our site has a
>>>> bunch
>>>>>> of products that are sold by any number of
>>>>>
>>>> sellers. Currently when I search
>>>>>> for some product I get back
>>>> all products
>>>>> matching that search term but the
>>>>>>
>>>> problem is there may be multiple products
>>>>> sold by the same seller
>>>> that are
>>>>>> all closely related, therefore their
>>>> scores
>>>>> are related. So basically the
>>>>>> search ends up
>>>> with results that are all
>>>>> closely clumped together by the same
>>>>
>>>>>> seller but I would much rather prefer
>>>>> to distribute
>>>> these results across
>>>>>> sellers (given each seller a fair shot
>>>> to
>>>>> sell their goods).
>>>>>
>>>>> Is there
>>>>
>>>>>> any way to add some boost query for example that will
>>>> start
>>>>> weighing products
>>>>>> lower when their seller has
>>>> already been listed a few
>>>>> times. For example,
>>>>>> right
>>>> now I have
>>>>>
>>>>> Product foo by Seller A
>>>>> Product
>>>> foo by Seller
>>>>>> A
>>>>> Product foo by Seller A
>>>>>
>>>> Product foo by Seller B
>>>>> Product foo by Seller
>>>>>>
>>>> B
>>>>> Product foo by Seller B
>>>>> Product foo by Seller
>>>> C
>>>>> Product foo by Seller
>>>>>> C
>>>>> Product foo
>>>> by Seller C
>>>>>
>>>>> where each result is very close in score. I
>>>>
>>>>>> would like something like this
>>>>>
>>>>> Product
>>>> foo by Seller A
>>>>> Product foo by
>>>>>> Seller B
>>>>>
>>>> Product foo by Seller C
>>>>> Product foo by Seller A
>>>>> Product
>>>> foo by
>>>>>> Seller B
>>>>> Product foo by Seller C
>>>>>
>>>> ....
>>>>>
>>>>> basically distributing the
>>>>>>
>>>> results over the sellers. Is something like this
>>>>> possible? I don't
>>>> care if
>>>>>> the solution involves a boost query or not. I
>>>> just
>>>>> want some way to
>>>>>> distribute closely related
>>>> documents.
>>>>>
>>>>> Thanks!!!
>>>>> --
>>>>> View
>>>> this
>>>>>> message in context:
>>>>>> href="
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>>
>>>>>> target=_blank
>>>>>> >
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>>>>
>>>> Sent
>>>>>> from the Solr - User mailing list archive at
>>>> Nabble.com.
>>>>>
>>>>>
>>>>
>>>>
>>>
>>> --
>>> View this
>>>> message in context:
>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
>>>> target=_blank
>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
>>> Sent
>>>> from the Solr - User mailing list archive at Nabble.com.
>>>
>>>
>>
>> --
>> View this message in context: http://old.nabble.com/Impossible-Boost-Query--tp28005354p28008495.html
>> Sent from the Solr - User mailing list archive at Nabble.com.
>>
>>
>
>
>
> --
> Lance Norskog
> goksron@gmail.com
>



--
Lance Norskog
goksron@gmail.com
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Lance Norskog-2
In reply to this post by Lance Norskog-2
The RandomValueSource class is available as a sort value, but it is
not available as a function. If it was, you could include the function
as part of the relevance but not all of it.

On Wed, Mar 24, 2010 at 9:41 AM, blargy <[hidden email]> wrote:

>
> This sound a little closer to what I want but I don't want fully randomized
> results.
>
> How exactly does this field work? Is it more than just a simple random sort
> (order by rand())? What would be nice is if I could randomize documents
> within a certain score percentage of each other. Is this available?
>
> Thanks
>
>
>
> Lance Norskog-2 wrote:
>>
>> Also, there is a 'random' type which generates random numbers. This
>> might help you also.
>>
>> On Tue, Mar 23, 2010 at 7:18 PM, Lance Norskog <[hidden email]> wrote:
>>> At this point (and for almost 3 years :) field collapsing is a source
>>> patch. You have to check out the Solr trunk from the Apache subversion
>>> server, apply the patch with the 'patch' command, and build the new
>>> Solr with 'ant'.
>>>
>>> On Tue, Mar 23, 2010 at 4:13 PM, blargy <[hidden email]> wrote:
>>>>
>>>> Thanks but Im not quite show on how to apply the patch. I just use the
>>>> packaged solr-1.4.0.war in my deployment (no compiling, etc). Is there a
>>>> way
>>>> I can patch the war file?
>>>>
>>>> Any instructions would be greatly appreciated. Thanks
>>>>
>>>>
>>>> Otis Gospodnetic wrote:
>>>>>
>>>>> You'd likely want to get the latest patch and trunk and try applying.
>>>>>
>>>>> Otis
>>>>> ----
>>>>> Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
>>>>> Hadoop ecosystem search :: http://search-hadoop.com/
>>>>>
>>>>>
>>>>>
>>>>> ----- Original Message ----
>>>>>> From: blargy <[hidden email]>
>>>>>> To: [hidden email]
>>>>>> Sent: Tue, March 23, 2010 6:10:22 PM
>>>>>> Subject: Re: Impossible Boost Query?
>>>>>>
>>>>>>
>>>>> Maybe a better question is... how can I install this and will it work
>>>>>> with
>>>>> 1.4?
>>>>>
>>>>> Thanks
>>>>>
>>>>>
>>>>> blargy wrote:
>>>>>>
>>>>>> Possibly.
>>>>>> How can I install this as a contrib or do I need to actually
>>>>>> perform the
>>>>>> patch?
>>>>>>
>>>>>>
>>>>>> Otis Gospodnetic wrote:
>>>>>>>
>>>>>>
>>>>>>> Would Field Collapsing from SOLR-236 do the job for
>>>>>> you?
>>>>>>>
>>>>>>> Otis
>>>>>>> ----
>>>>>>> Sematext ::
>>>>>> href="http://sematext.com/" target=_blank >http://sematext.com/ ::
>>>>>> Solr -
>>>>>> Lucene - Nutch
>>>>>>> Hadoop ecosystem search ::
>>>>>> href="http://search-hadoop.com/" target=_blank
>>>>>> >http://search-hadoop.com/
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>> ----- Original Message ----
>>>>>>>> From: blargy <
>>>>>> ymailto="mailto:[hidden email]"
>>>>>> href="mailto:[hidden email]">[hidden email]>
>>>>>>>>
>>>>>> To:
>>>>>> href="mailto:[hidden email]">[hidden email]
>>>>>>>>
>>>>>> Sent: Tue, March 23, 2010 2:39:48 PM
>>>>>>>> Subject: Impossible Boost
>>>>>> Query?
>>>>>>>>
>>>>>>>>
>>>>>>> I was wondering if this is
>>>>>> even possible. I'll try to explain what I'm
>>>>>>>> trying
>>>>>>>
>>>>>> to do to the best of my ability.
>>>>>>>
>>>>>>> Ok, so our site has a
>>>>>> bunch
>>>>>>>> of products that are sold by any number of
>>>>>>>
>>>>>> sellers. Currently when I search
>>>>>>>> for some product I get back
>>>>>> all products
>>>>>>> matching that search term but the
>>>>>>>>
>>>>>> problem is there may be multiple products
>>>>>>> sold by the same seller
>>>>>> that are
>>>>>>>> all closely related, therefore their
>>>>>> scores
>>>>>>> are related. So basically the
>>>>>>>> search ends up
>>>>>> with results that are all
>>>>>>> closely clumped together by the same
>>>>>>
>>>>>>>> seller but I would much rather prefer
>>>>>>> to distribute
>>>>>> these results across
>>>>>>>> sellers (given each seller a fair shot
>>>>>> to
>>>>>>> sell their goods).
>>>>>>>
>>>>>>> Is there
>>>>>>
>>>>>>>> any way to add some boost query for example that will
>>>>>> start
>>>>>>> weighing products
>>>>>>>> lower when their seller has
>>>>>> already been listed a few
>>>>>>> times. For example,
>>>>>>>> right
>>>>>> now I have
>>>>>>>
>>>>>>> Product foo by Seller A
>>>>>>> Product
>>>>>> foo by Seller
>>>>>>>> A
>>>>>>> Product foo by Seller A
>>>>>>>
>>>>>> Product foo by Seller B
>>>>>>> Product foo by Seller
>>>>>>>>
>>>>>> B
>>>>>>> Product foo by Seller B
>>>>>>> Product foo by Seller
>>>>>> C
>>>>>>> Product foo by Seller
>>>>>>>> C
>>>>>>> Product foo
>>>>>> by Seller C
>>>>>>>
>>>>>>> where each result is very close in score. I
>>>>>>
>>>>>>>> would like something like this
>>>>>>>
>>>>>>> Product
>>>>>> foo by Seller A
>>>>>>> Product foo by
>>>>>>>> Seller B
>>>>>>>
>>>>>> Product foo by Seller C
>>>>>>> Product foo by Seller A
>>>>>>> Product
>>>>>> foo by
>>>>>>>> Seller B
>>>>>>> Product foo by Seller C
>>>>>>>
>>>>>> ....
>>>>>>>
>>>>>>> basically distributing the
>>>>>>>>
>>>>>> results over the sellers. Is something like this
>>>>>>> possible? I don't
>>>>>> care if
>>>>>>>> the solution involves a boost query or not. I
>>>>>> just
>>>>>>> want some way to
>>>>>>>> distribute closely related
>>>>>> documents.
>>>>>>>
>>>>>>> Thanks!!!
>>>>>>> --
>>>>>>> View
>>>>>> this
>>>>>>>> message in context:
>>>>>>>> href="
>>>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>>>> target=_blank
>>>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>>>>
>>>>>>>> target=_blank
>>>>>>>> >
>>>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html"
>>>>>> target=_blank
>>>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28005354.html
>>>>>>>
>>>>>> Sent
>>>>>>>> from the Solr - User mailing list archive at
>>>>>> Nabble.com.
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>> --
>>>>> View this
>>>>>> message in context:
>>>>>> href="http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html"
>>>>>> target=_blank
>>>>>> >http://old.nabble.com/Impossible-Boost-Query--tp28005354p28007880.html
>>>>> Sent
>>>>>> from the Solr - User mailing list archive at Nabble.com.
>>>>>
>>>>>
>>>>
>>>> --
>>>> View this message in context:
>>>> http://old.nabble.com/Impossible-Boost-Query--tp28005354p28008495.html
>>>> Sent from the Solr - User mailing list archive at Nabble.com.
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Lance Norskog
>>> [hidden email]
>>>
>>
>>
>>
>> --
>> Lance Norskog
>> [hidden email]
>>
>>
>
> --
> View this message in context: http://old.nabble.com/Impossible-Boost-Query--tp28005354p28017469.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
>



--
Lance Norskog
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

Blargy
Ok so this is basically just a random sort.

Anyway I can get this to randomly sort documents that closely related and not the rest of the results?
Reply | Threaded
Open this post in threaded view
|

Re: Impossible Boost Query?

britske
Have a look at functionqueries.

http://wiki.apache.org/solr/FunctionQuery

<http://wiki.apache.org/solr/FunctionQuery>You could for instance use your
regular score and multiply it with RandomValueSource bound between 1.0 and
1.1 for example.
This would at least break ties in a possibly natural looking manner.  (btw:
this would still influence all documents however)

//Geert-Jan

2010/3/26 Blargy <[hidden email]>

>
> Ok so this is basically just a random sort.
>
> Anyway I can get this to randomly sort documents that closely related and
> not the rest of the results?
> --
> View this message in context:
> http://n3.nabble.com/Impossible-Boost-Query-tp472080p580214.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>