The swarm of bees!

The Spark SQL parser is limited in terms of the queries you can write, and so SparkSQL provides a Hive Context, with HiveQL parser. Calliope also supports HiveContext to execute HiveQL on Cassandra. All the functionality provided by HiveContext (including the JDBC Server) is present in Calliope's HiveContext too.

The package

To use HiveQl with Cassandra from the Spark Shell, add this assembly jar to spark-shell's and executor's classpath.

$ bin/spark-shell --jars calliope-sql-assembly-1.1.0-CTP-U2.jar

Just like Calliope SQL, you need to add calliope-hive to your project's dependencies.

Maven:

<dependency>
    <groupId>com.tuplejump</groupId>
    <artifactId>calliope-hive_2.10</artifactId>
    <version>1.1.0-CTP-U2</version>
</dependency>

SBT:

libraryDependencies += "com.tuplejump" % "calliope-hive_2.10" % "1.1.0-CTP-U2"

Creating the Hive Context

Like the CassandraAwareSQLContext we have a CassandraAwareHiveContext, which extends the SparkSQL HiveContext to provide Cassandra relevant functionality.

val sc: SparkContext // An existing SparkContext.
val sqlContext = new org.apache.spark.sql.hive.CassandraAwareHiveContext(sc)

import sqlContext.createSchemaRDD

Now you have a HiveContext that you can use same as documented in the Spark SQL Programming Guide

Additionally, this HiveContext also knows about the configured (see sql config) Cassandra cluster and is all set to run your Hive Queries against it.

This, also provides the same programming API as Calliope's SQL Context, so you can create a SchemaRDD programmatically.