This library lets you expose Cassandra tables as Spark RDDs, write Spark RDDs to Cassandra tables, and execute arbitrary CQL queries in your Spark applications.
- Compatible with Apache Cassandra version 2.0 or higher and DataStax Enterprise 4.5 (see table below)
- Compatible with Apache Spark 1.0 and 1.1 (see table below)
- Exposes Cassandra tables as Spark RDDs
- Maps table rows to CassandraRow objects or tuples
- Offers customizable object mapper for mapping rows to objects of user-defined classes
- Saves RDDs back to Cassandra by implicit
saveToCassandracall - Converts data types between Cassandra and Scala
- Supports all Cassandra data types including collections
- Filters rows on the server side via the CQL
WHEREclause - Allows for execution of arbitrary CQL statements
- Plays nice with Cassandra Virtual Nodes
The connector project has several branches, each of which map into different supported versions of Spark and Cassandra. Refer to the compatibility table below which shows the major.minor version range supported between the connector, Spark, Cassandra, and the Cassandra Java driver:
| Connector | Spark | Cassandra | Cassandra Java Driver |
|---|---|---|---|
| 1.2 | 1.2, 1.1, 1.0 | 2.1, 2.0 | 2.1 |
| 1.1 | 1.1, 1.0 | 2.1, 2.0 | 2.1 |
| 1.0 | 1.0, 0.9 | 2.0 | 1.0 |
This project has been published to the Maven Central Repository. For SBT to download the connector binaries, sources and javadoc, put this in your project SBT config:
libraryDependencies += "com.datastax.spark" %% "spark-cassandra-connector" % "1.2.0-alpha1"
If you want to access the functionality of Connector from Java, you may want to add also a Java API module:
libraryDependencies += "com.datastax.spark" %% "spark-cassandra-connector-java" % "1.2.0-alpha1"
In the root directory run
sbt assembly
A fat jar will be generated to both of these directories:
spark-cassandra-connector/target/scala-2.10/spark-cassandra-connector-java/target/scala-2.10/
Select the former for Scala apps, the later for Java.
In the root directory run:
sbt package
sbt doc
The library package jars will be placed in:
spark-cassandra-connector/target/scala-2.10/spark-cassandra-connector-java/target/scala-2.10/
The documentation will be generated to:
spark-cassandra-connector/target/scala-2.10/api/spark-cassandra-connector-java/target/scala-2.10/api/
- Quick-start guide
- Connecting to Cassandra
- Loading datasets from Cassandra
- Server-side data selection and filtering
- Working with user-defined case classes and tuples
- Saving datasets to Cassandra
- Customizing the object mapping
- Using Connector in Java
- Spark Streaming with Cassandra
- About The Demos
- The spark-cassandra-connector-embedded Artifact
This software is available under the Apache License, Version 2.0.
New issues should be reported using JIRA. Please do not use the built-in GitHub issue tracker. It is left for archival purposes and it will be disabled soon.
Questions etc can be submitted to the user mailing list.
To develop this project, we recommend using IntelliJ IDEA. Make sure you have installed and enabled the Scala Plugin. Open the project with IntelliJ IDEA and it will automatically create the project structure from the provided SBT configuration.
Before contributing your changes to the project, please make sure that all unit tests and integration tests pass. Don't forget to add an appropriate entry at the top of CHANGES.txt. Finally open a pull-request on GitHub and await review.
If your pull-request is going to resolve some opened issue, please add Fixes #xx at the end of each commit message (where xx is the number of the issue).
To run unit and integration tests:
./sbt/sbt test
./sbt/sbt it:test
By default, integration tests start up a separate, single Cassandra instance and run Spark in local mode. It is possible to run integration tests with your own Cassandra and/or Spark cluster. First, prepare a jar with testing code:
./sbt/sbt test:package
Then copy the generated test jar to your Spark nodes and run:
export IT_TEST_CASSANDRA_HOST=<IP of one of the Cassandra nodes>
export IT_TEST_SPARK_MASTER=<Spark Master URL>
./sbt/sbt it:test