Using Apache Hive
Also available as:

Set up the development environment

You can create a Hive UDF in a development environment using IntelliJ, for example, and build the UDF with Hive and Hadoop JARS that you download from your HDP 3.x cluster.

  1. On your cluster, locate the hadoop-common-<version>.jar and hive-exec-<version>.jar.
    For example:
    ls /usr/hdp/current/hadoop-client/hadoop-common-* |grep -v test
    # ls /usr/hdp/current/hive-server2/lib/hive-exec-*
  2. Download the JARs to your development computer to add to your IntelliJ project later.
  3. Open IntelliJ and create a new Maven-based project. Click Create New Project, select Maven, and select Java version 1.8 as the Project SDK. Click Next.
  4. Add archetype information.
    For example:
    • GroupId: com.mycompany.hiveudf
    • ArtifactId: hiveudf
  5. Click Next and Finish.
    The generated pom.xml appears in sample-hiveudf.
  6. To the pom.xml, add properties to facilitate versioning.
    For example:
  7. In the pom.xml, define the repositories.
    Use internal repositories if you do not have internet access.
          <name>HDP Releases</name>
          <name>Public Hortonworks Maven Repo</name>
  8. Define dependencies.
    For example:
  9. Select File > Project Structure. Click Modules. On the Dependencies tab, click + to add JARS or directories. Browse to and select the JARs you downloaded in step 1.