AWS Redshift driver in Zeppelin

Question

I want to explore my data in Redshift using notebook Zeppelin. A small EMR cluster with Spark is running behind. I am loading databricks' spark-redshift library

%dep
z.reset()
z.load("com.databricks:spark-redshift_2.10:0.6.0")

and then

import org.apache.spark.sql.DataFrame

val query = "..."

val url = "..."
val port=5439
val table = "..."
val database = "..."
val user = "..."
val password = "..."

val df: DataFrame = sqlContext.read
  .format("com.databricks.spark.redshift")
  .option("url", s"jdbc:redshift://${url}:$port/$database?user=$user&password=$password")
  .option("query",query)
  .option("tempdir", "s3n://.../tmp/data")
  .load()

df.show

but I get the error

java.lang.ClassNotFoundException: Could not load an Amazon Redshift JDBC driver; see the README for instructions on downloading and configuring the official Amazon driver

I added option

option("jdbcdriver", "com.amazon.redshift.jdbc41.Driver")

but not for the better. I think I need to specify redshift's JDBC driver somewhere like I would passing --driver-class-path to spark-shell, but how to do that with zeppelin?

Answer 1

You can add external jars with dependencies like the JDBC driver using either Zeppelin's dependency-loading mechanism or, in case of Spark, using %dep dynamic dependency loader

When your code requires external library, instead of doing download/copy/restart Zeppelin, you can easily do following jobs using %dep interpreter.

• Load libraries recursively from Maven repository
• Load libraries from local filesystem
• Add additional maven repository
• Automatically add libraries to SparkCluster (You can turn off)

The latter would look something like:

%dep
// loads with all transitive dependencies from Maven repo
z.load("groupId:artifactId:version")

// or add artifact from filesystem
z.load("/path/to.jar")

and by convention have to be in the first paragraph of the note.