Compiling the Samples on Windows
When you build the samples, the tests might not run on Windows without additional configuration. This topic provides a workaround procedure.
Procedure
-
In the
PluginExampleProject, run the following command:
mvn package -DskipTests
The project should build correctly and provide you with the JAR necessary to upload into Team Studio.
If you want to be able to run the tests, be award that Spark has extra requirements for reading and saving data in a Spark context on Windows. The linear regression model from MLlib (unlike the other Spark operators in PluginExampleProject) uses some Hadoop functionality to save intermediate results. Running Hadoop on Windows requires that you have a local Hadoop installation, the HADOOP_HOME variable correctly configured, and the winutils executable present on your computer. You might experience build errors if you are testing functionality using Spark Core or a Hive context. This is particularly true if you want to read and save data in a Windows environment. For details, see SPARK-2356 and How to run Apache Spark on Windows7 in standalone mode