Two probably [hopefully] basic questions

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

Two probably [hopefully] basic questions

Thomas Dowling-2
I have some sort of blind spot that's preventing me from seeing this in
the online documentation, but how do I know if I'm using the standard or
dismax request handlers, and how do I change between them?

Also, is there something about boosting fields at index time that messes
with the relevance of phrase searching?  Let me explain: I have one
index of ~280k bibliographic records, with no field boosting.  If I
search for HABITAT LOSS, the top hits all have the two-word phrase
"habitat loss".  I have another index of about ~840k bibliographic
records, with titles and abstracts boosted slightly upwards; when I
search HABITAT LOSS my top hits all have "habitat" or of "loss" by
themselves in the abstract (there are hits on the phrase "habitat loss"
in the abstract and I can find them with a quoted search string).

All in Solr 1.2 with no substantial differences between their schema.xml
files.

--
Thomas Dowling
[hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: Two probably [hopefully] basic questions

Yonik Seeley-2
On Thu, Feb 28, 2008 at 11:44 AM, Thomas Dowling <[hidden email]> wrote:
> I have some sort of blind spot that's preventing me from seeing this in
>  the online documentation, but how do I know if I'm using the standard or
>  dismax request handlers, and how do I change between them?

qt specifies the handler (originally stood for query type).  It
defaults to standard.

In Solr 1.3, the dismax handler is actually now just the standard
handler with a default of defType=dismax (defType stands for default
query type, and changes what query parser is invoked).

>  Also, is there something about boosting fields at index time that messes
>  with the relevance of phrase searching?

Not to my knowledge.   Are you sure in the latter case below that you
are using dismax and are adding phrases via pf?  If so, perhaps just
try boosting the phrase searches more to compensate for the fact that
some documents are boosted.

-Yonik

>  Let me explain: I have one
>  index of ~280k bibliographic records, with no field boosting.  If I
>  search for HABITAT LOSS, the top hits all have the two-word phrase
>  "habitat loss".  I have another index of about ~840k bibliographic
>  records, with titles and abstracts boosted slightly upwards; when I
>  search HABITAT LOSS my top hits all have "habitat" or of "loss" by
>  themselves in the abstract (there are hits on the phrase "habitat loss"
>  in the abstract and I can find them with a quoted search string).
>
>  All in Solr 1.2 with no substantial differences between their schema.xml
>  files.
>
>  --
>  Thomas Dowling
>  [hidden email]
>
>