> We have documents which are comprised
> of:
> 
>     - A short list of terms (about 1 to 5 terms
> per document)
>     - An estimate of the probability of the terms
> occurrence (stored as tint)
> 
> For each term in the index, we would like to get the result
> of the following function:
> 
>     (our estimate of the probability/100) x (a
> term's Document Frequency)
> 
> So if the term "fox" occurred in 7 documents, the desired
> query result would look something like:
> 
> <doc>
> <str name="term">fox</str>
> <str name="DF">7</str>
> <int name="probability">23</int>
> <str name="functionResult">1.61</str>
> </doc>
> 
> We can find a number of examples for using function queries
> to alter scoring or sorting results, but can not find any
> that show how to get the value of actual function result
> back.


It seems that your requirement is not related to function queries. You are 
interested something that is term level rather than document level.

Anyway with http://wiki.apache.org/solr/TermsComponent you can iterate through 
all terms and it gives you DF information. You can multiple that DF with 
pre-computed (term level) probability. To get all terms use -1 as limit. e.g. = 
&terms.limit=-1



Reply via email to