excludeIds in QueryElevationComponent (4.7)

[Lajos <la...@protulae.com>]

Guys,

I've been testing out https://issues.apache.org/jira/browse/SOLR-5541 on 
4.7RC4.

I previously had an elevate.xml that elevated 3 documents for a specific 
query. My understanding is that I could, at runtime, exclude one of 
those. So I tried that like this:

http://localhost:8080/solr/ecommerce/search?q=canon&excludeIds=208464207

and now NONE of my documents are elevated. What I would have expected is 
that I'd have 2 elevated documents, but the 208464207 would not be 
amongst them.

Sadly, what happens is that now nothing is elevated.

Am I misunderstanding something or should I open a JIRA? Looking at the 
source code I can't immediately see what would be wrong.

Thanks,

Lajos

Re: excludeIds in QueryElevationComponent (4.7)

[Lajos <la...@protulae.com>]

Thanks Hoss, that makes sense.

Anyway, I like the new paradigm better ... it allows for more 
intelligent elevation control.

Cheers,

L


On 25/02/2014 23:26, Chris Hostetter wrote:
>
> : What is seems that is happening is that excludeIds or elevateIds ignores
> : what's in elevate.xml. I would have expected (hoped) that it would layer on
> : top of that, which makes a bit more sense I think.
>
> That's not how it's implemented -- i believe Joel implemented this way
> intentional because otherwise if the elevate.xml said "elevate A,B and
> exclude X,Y" there would be no simple way to say "instead of what's in
> elevate.xml, i want to elevate X,Y and i don't wnat to exclude *anything*"
>
> I made sure this was explicitly documented in the ref guide...
>
> https://cwiki.apache.org/confluence/display/solr/The+Query+Elevation+Component#TheQueryElevationComponent-TheelevateIdsandexcludeIdsParameters
>
> "If either one of these parameters is specified at request time, the the
> entire elevation configuration for the query is ignored."
>
>
>
> -Hoss
> http://www.lucidworks.com/
>

Re: excludeIds in QueryElevationComponent (4.7)

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

: What is seems that is happening is that excludeIds or elevateIds ignores
: what's in elevate.xml. I would have expected (hoped) that it would layer on
: top of that, which makes a bit more sense I think.

That's not how it's implemented -- i believe Joel implemented this way 
intentional because otherwise if the elevate.xml said "elevate A,B and 
exclude X,Y" there would be no simple way to say "instead of what's in 
elevate.xml, i want to elevate X,Y and i don't wnat to exclude *anything*"

I made sure this was explicitly documented in the ref guide...

https://cwiki.apache.org/confluence/display/solr/The+Query+Elevation+Component#TheQueryElevationComponent-TheelevateIdsandexcludeIdsParameters

"If either one of these parameters is specified at request time, the the 
entire elevation configuration for the query is ignored."



-Hoss
http://www.lucidworks.com/

Re: excludeIds in QueryElevationComponent (4.7)

[Lajos <la...@protulae.com>]

Hit the send button too fast ...

What is seems that is happening is that excludeIds or elevateIds ignores 
what's in elevate.xml. I would have expected (hoped) that it would layer 
on top of that, which makes a bit more sense I think.

Thanks,

Lajos


On 25/02/2014 22:58, Lajos wrote:
> Guys,
>
> I've been testing out https://issues.apache.org/jira/browse/SOLR-5541 on
> 4.7RC4.
>
> I previously had an elevate.xml that elevated 3 documents for a specific
> query. My understanding is that I could, at runtime, exclude one of
> those. So I tried that like this:
>
> http://localhost:8080/solr/ecommerce/search?q=canon&excludeIds=208464207
>
> and now NONE of my documents are elevated. What I would have expected is
> that I'd have 2 elevated documents, but the 208464207 would not be
> amongst them.
>
> Sadly, what happens is that now nothing is elevated.
>
> Am I misunderstanding something or should I open a JIRA? Looking at the
> source code I can't immediately see what would be wrong.
>
> Thanks,
>
> Lajos