On Fri, Jul 24, 2015, at 12:53 AM, Rex X wrote: > Given following Impala query: > > SELECT date, SUM(CAST(price AS DOUBLE)) AS price > FROM table > WHERE date='2014-01-01' AND store_id IN(1,2,3) > GROUP BY date; > > To work with Solr > > 1. Will it be more efficient to directly use equivalent Solr query? Any > curl command equivalent to the Impala query above? Or > 2. Will it be faster to create a new table based on the query above with > Impala, and then connect Impala with Solr? Any such Impala-Solr > connector? > > The final goal is to use Kibana to connect Solr for visualization. > > Any comments are greatly welcome!
I do not know Impala so cannot comment much on that - i.e. would querying Solr or Impala be more efficient? No idea. The above looks like an aggregation with filtering, so I'd suggest you look at the new json facet API in Solr which would get your aggregations (and summing). To query against Solr, you need to have pushed your content *to* Solr. It won't go ask Impala for you. You will have to set up mechanisms for your content to get into Solr for Solr to be any use. Lastly, Kibana is a tool that works on top of Elasticsearch. To use Solr, you should look at Lucidworks Banana in its place. Upayavira
