This is answered a few times already here. Have a look at for example http://community.cloudera.com/t5/Advanced-Analytics-Apache-Spark/I-am-using-a-hive-cotext-in-pyspark...

The short answer is that Spark is not entirely compatible with recent versions of Hive found in CDH, but may still work for a lot of use cases. The Spark bits are still there. You have to add Hive to the classpath yourself.