New Contributor
Posts: 1
Registered: ‎05-08-2017

Impala odbc performance with spss modeler

Hi folks,


We are trying to test impala in our BI Team after recommandations from cloudera. Our hadoop cluster is a cdh 5.7.0.

We could connect to the metastore in hdfs with our BI tool through the odbc driver , but, when running a query:

Expected behaviour:

  • The query should be processed on the server side(the hadoop cluster)
  • Only the needed resulst should sent to our BI Tool

Actual behaviour:

  • All the concerned tables are sent to the client
  • The query is processed on the client side(really?, or did we miss something)

We use the last odbc driver version 2.5.37.


What could be that we are not doing right?

Is there some recommandations when working using impala with SPSS Modeller? Other driver (JDBC or ODBC ?, from cloudera or someone else?)