Query Cassandra from Spark using CassandraSQLContext

220 views Asked by At

I try to query Cassandra from Spark using CassandraSQLContext, but I get an weird missing dependency error. I have a Spark application like the following :

val spark: SparkSession = SparkSession.builder().appName(appName).getOrCreate()
val cassandraSQLContext = new org.apache.spark.sql.cassandra.CassandraSQLContext(spark.sparkContext)

val path = args(0)
cassandraSQLContext.setKeyspace(args(1))

val dataFrame: DataFrame = cassandraSQLContext.sql(args(2))
dataFrame.write.mode(SaveMode.Overwrite).option("header", "true").csv(path)

And I get a missing Spark scala class error :

User class threw exception: java.lang.NoClassDefFoundError: org/apache/spark/sql/catalyst/analysis/Catalog
    at com.test.batch.utils.CSVFromCassandraSQLQuery$.<init>(CSVFromCassandraSQLQuery.scala:19)
    at com.test.batch.utils.CSVFromCassandraSQLQuery$.<clinit>(CSVFromCassandraSQLQuery.scala)
    at com.test.batch.utils.CSVFromCassandraSQLQuery.main(CSVFromCassandraSQLQuery.scala)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at org.apache.spark.deploy.yarn.ApplicationMaster$$anon$4.run(ApplicationMaster.scala:721)
Caused by: java.lang.ClassNotFoundException: org.apache.spark.sql.catalyst.analysis.Catalog
    at java.net.URLClassLoader.findClass(URLClassLoader.java:381)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:424)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:357)
    ... 8 more

I also tried to add the spark-catalyst jar path explicitly in the spark-submit command, but I still get the same issue (whether I run it locally or on a yarn cluster) ...

Here is my project settings build.sbt :

scalaVersion := "2.11.11"
val sparkVersion = "2.3.1"

libraryDependencies ++= Seq(
  "log4j" % "log4j" % "1.2.17",
  "org.apache.spark" %% "spark-core" % sparkVersion % "provided",
  "org.apache.spark" %% "spark-sql" % sparkVersion  % "provided",
  "org.apache.spark" %% "spark-hive" % sparkVersion % "provided",
  "org.apache.spark" %% "spark-catalyst" % sparkVersion % "provided",
  "org.apache.spark" %% "spark-streaming" % sparkVersion % "provided",
  "com.datastax.spark" %% "spark-cassandra-connector" % "2.3.1",
  "org.scala-lang.modules" %% "scala-java8-compat" % "0.9.0",
  "com.twitter" % "jsr166e" % "1.1.0"
)

Any ideas about what I'm missing ?

1

There are 1 answers

0
KZapagol On BEST ANSWER

This issue is occuring because org.apache.spark.sql.catalyst.analysis.Catalog class is not part of spark-catalyst_2.11:2.x.x anymore. Please refer source code of spark-catalyst_2.11:2.0.0

https://jar-download.com/artifacts/org.apache.spark/spark-catalyst_2.11/2.0.0/source-code

org.apache.spark.sql.catalyst.analysis.Catalog class is available till spark-catalyst_2.11:1.6.3 version.

https://jar-download.com/artifacts/org.apache.spark/spark-catalyst_2.11/1.6.3/source-code

I would request you to not use CassandraSQLContext as it has been deprecated. Please check https://datastax-oss.atlassian.net/browse/SPARKC-399.

Please check SO post to use Cassandra Context in spark 2.0.

how to use Cassandra Context in spark 2.0