Cloudera document says "Cloudera recommends using the Cloudera JDBC Connector where practical"
What are the advantages of Cloudera JDBC Connector over Hive jdbc driver for impala? why does cloudera recommend cloudera jdbc connector?
We don't officially support the Hive JDBC driver with Impala so we don't really guarantee that it will work in all cases, e.g. especially with the various secure communication and authentication options.
The Cloudera JDBC driver also implements various features client-side like substitution variables. It also does some query rewrites to support additional SQL constructs if you have "UseNativeQuery=0", although Impala's SQL support is a lot more complete now so I'm not sure how many cases this still applies to.