Skip to main content
Version: 1.37.0

Installation

To integrate OpenLineage Spark with your application, you can:

Bundle the package with your Apache Spark application project​

info

This approach does not demonstrate how to configure the OpenLineageSparkListener. Please refer to the Configuration section.

For Maven, add the following to your pom.xml:

<dependency>
  <groupId>io.openlineage</groupId>
  <artifactId>openlineage-spark_${SCALA_BINARY_VERSION}</artifactId>
  <version>1.39.0</version>
</dependency>

For Gradle, add this to your build.gradle:

implementation("io.openlineage:openlineage-spark_${SCALA_BINARY_VERSION}:1.39.0")

Place the JAR in your ${SPARK_HOME}/jars directory​

info

This approach does not demonstrate how to configure the OpenLineageSparkListener. Please refer to the Configuration section.

  1. Download the JAR and its checksum from Maven Central.
  2. Verify the JAR's integrity using the checksum.
  3. Upon successful verification, move the JAR to ${SPARK_HOME}/jars.

This script automates the download and verification process:

#!/usr/bin/env bash

if [ -z "$SPARK_HOME" ]; then
    echo "SPARK_HOME is not set. Please define it as your Spark installation directory."
    exit 1
fi

OPENLINEAGE_SPARK_VERSION='1.39.0'
SCALA_BINARY_VERSION='2.13'        # Example Scala version
ARTIFACT_ID="openlineage-spark_${SCALA_BINARY_VERSION}"
JAR_NAME="${ARTIFACT_ID}-${OPENLINEAGE_SPARK_VERSION}.jar"
CHECKSUM_NAME="${JAR_NAME}.sha512"
BASE_URL="https://repo1.maven.org/maven2/io/openlineage/${ARTIFACT_ID}/${OPENLINEAGE_SPARK_VERSION}"

curl -O "${BASE_URL}/${JAR_NAME}"
curl -O "${BASE_URL}/${CHECKSUM_NAME}"

echo "$(cat ${CHECKSUM_NAME})  ${JAR_NAME}" | sha512sum -c

if [ $? -eq 0 ]; then
    mv "${JAR_NAME}" "${SPARK_HOME}/jars"
else
    echo "Checksum verification failed."
    exit 1
fi

Use the --jars option with spark-submit / spark-shell / pyspark​

info

This approach does not demonstrate how to configure the OpenLineageSparkListener. Please refer to the Configuration section.

  1. Download the JAR and its checksum from Maven Central.
  2. Verify the JAR's integrity using the checksum.
  3. Upon successful verification, submit a Spark application with the JAR using the --jars option.

This script demonstrate this process:

#!/usr/bin/env bash

OPENLINEAGE_SPARK_VERSION='1.39.0'
SCALA_BINARY_VERSION='2.13'        # Example Scala version
ARTIFACT_ID="openlineage-spark_${SCALA_BINARY_VERSION}"
JAR_NAME="${ARTIFACT_ID}-${OPENLINEAGE_SPARK_VERSION}.jar"
CHECKSUM_NAME="${JAR_NAME}.sha512"
BASE_URL="https://repo1.maven.org/maven2/io/openlineage/${ARTIFACT_ID}/${OPENLINEAGE_SPARK_VERSION}"

curl -O "${BASE_URL}/${JAR_NAME}"
curl -O "${BASE_URL}/${CHECKSUM_NAME}"

echo "$(cat ${CHECKSUM_NAME})  ${JAR_NAME}" | sha512sum -c

if [ $? -eq 0 ]; then
    spark-submit --jars "path/to/${JAR_NAME}" \
      # ... other options
else
    echo "Checksum verification failed."
    exit 1
fi

Use the --packages option with spark-submit / spark-shell / pyspark​

info

This approach does not demonstrate how to configure the OpenLineageSparkListener. Please refer to the Configuration section.

Spark allows you to add packages at runtime using the --packages option with spark-submit. This option automatically downloads the package from Maven Central (or other configured repositories) during runtime and adds it to the classpath of your Spark application.

OPENLINEAGE_SPARK_VERSION='1.39.0'
SCALA_BINARY_VERSION='2.13'        # Example Scala version

spark-submit --packages "io.openlineage:openlineage-spark_${SCALA_BINARY_VERSION}:1.39.0" \
    # ... other options
warning

Version 1.8.0 and earlier only supported Scala 2.12 variants of Apache Spark. Scala version name was not included in the artifact identifier.