Yep. It's almost always easier and faster if you can pre-compute as much as possible during indexing time. It'll take longer to index of course, but the ratio of writing to the index to searching is usually hugely in favor of doing the work during indexing.
Best, Erick On Mon, Nov 3, 2014 at 8:52 PM, Yubing (Tom) Dong 董玉冰 <tom.tung....@gmail.com> wrote: > Hi Erik, > > Thanks for the reply! Do you mean parse and modify the documents before > sending them to Solr? > > Cheers, > Yubing > > On Mon, Nov 3, 2014 at 8:48 PM, Erick Erickson <erickerick...@gmail.com> > wrote: > >> Wouldn't it be easiest to compute the span at index time? Then it's >> very straight-forward. >> >> Best, >> Erick >> >> On Mon, Nov 3, 2014 at 8:18 PM, Yubing (Tom) Dong 董玉冰 >> <tom.tung....@gmail.com> wrote: >> > Hi, >> > >> > I'm new to Solr, and I'm having a problem with faceting. I would really >> > appreciate it if you could help :) >> > >> > I have a set of documents in JSON format, which I could post to my Solr >> > core using the post.jar tool. Each document contains two fields, namely >> > "startDate" and "endDate", both of which are of type "date". >> > >> > Conceptually, I would like to have a third field "timeSpan" that is >> > automatically generated from the return value of function query >> > "ms(endDate, startDate)", and do range facet on it, i.e. compute the >> > distribution of "timeSpan", among either all of or a filtered subset of >> the >> > documents. >> > >> > I have tried to find ways of both directly faceting the function return >> > values and automatically generate the "timeSpan" field during indexing, >> but >> > without luck yet. >> > >> > Suggestions are greatly appreciated! >> > >> > Best, >> > Yubing >>
