We have an updated version of Knime and Spark (1.7).
When I execute Spark Decision Tree Learner node, I am getting following error. Not sure why it says incompatible.
Any insights into this would be appreciated.
"ERROR Spark Decision Tree Learner 3:235 Execute failed: An error occured: org.apache.spark.mllib.tree.model.DecisionTreeModel
technically this error means that KNIME AP (the client) cannot read the Decision Tree model that is coming back from the Decision Tree learner Spark job. Usually that indicates that the version of Spark you have set in KNIME Analytics Platform is different from the one in your cluster.
You can set the version of Spark your KNIME Analytics Platform uses under File > Preferences > Big Data Extensions > Spark. It must match the one installed in your cluster.
Out of curiosity: Which Hadoop distribution & version (example: Cloudera 5.8) with which version of Spark (e.g. Spark 1.6) are you using on the cluster side?