I think the StatsComponent will do everything you're asking for here:
http://wiki.apache.org/solr/StatsComponent
On Jan 8, 2010, at 9:37 AM, dbashford wrote:
>
> I get the feeling what I need to accomplish isn't necessarily in the
> spirit
> of what solr is meant to do, but it's the problem I'm facing. Of
> course,
> I'm a solr newbie, so this may not be as challenging as I think it is.
>
> Domain is a little tricky, so I'll make one up. Lets say I have the
> following 3 pieces of information in my index ==> Book Id, Book
> Category(mystery, fantasy, etc), # of Pages. And lets say there's a
> million, two million...a lot of books.
>
> Question 1) Is there a way to get a total number of pages across all
> books
> without pulling every book and iterating over the result? (Lets
> pretend
> doing so is a worthwhile endeavor.)
>
> One thing I've found I can do to reduce iteration is to do a term
> search on
> # of pages, which does a little gathering up for me. If books have
> between
> 100 and 1000 pages, I've got 900 results to iterate over, little
> multiplying
> and adding and I have what I need. That's acceptable, and in this
> case its
> definitely faster than going to the database and querying a poorly
> normalized group of tables, but still, if there was some way to just
> get a
> sum...
>
> Question 2) The next question would be, assuming there's some way to
> get a
> sum for all books, how would I get a sum for all books that are
> mysteries?
> I don't think the term search works here because I can't seem to
> find a way
> to limit the docs the term search uses.
>
> Thanks in advance for any help you can provide!
>
> --
> View this message in context: http://old.nabble.com/Sum---Filter-in-term-search---Something-else--tp27076924p27076924.html
> Sent from the Solr - User mailing list archive at Nabble.com.
>
|