Hello, This sounds like a custom SearchComponent. Which clustering library you want to use or DIY is up to you, but go with the SearchComponent approach. You will still need to process N hits, but you won't need to first send them all over the wire.
Otis -- Solr & ElasticSearch Support http://sematext.com/ On Fri, Jun 7, 2013 at 11:48 AM, Raheel Hasan <raheelhasan....@gmail.com> wrote: > Hi, > > Can someone please tell me if there is a way to have a custom *`clustering > of the data`* from `solr` 'query' results? I am facing 2 issues currently: > > 1. The `*Carrot*` clustering only applies clustering to the "paged" > results (i.e. in the current pagination's page results). > > 2. I need to have custom clustering and classify results into certain > classes only (i.e. only few very specific words in the search results). > Like for example "Red", "Green", "Blue" etc... and not "hello World", > "Known World", "green world" etc.... -(if you know what I mean here) - > Where all these words in both Do and DoNot existing in the search results. > > Please tell me how to achieve this. Perhaps Carrot/clustering is not needed > here and some other classifier is needed. So what to do here? > > Basically, I cannot receive 1 million results, then process them via > PHP-Array to classify them as per need. The classification must be done > here in solr only. > > Thanks > > -- > Regards, > Raheel Hasan
