Answer from Java side is empty - MMLspark

Hi all,

I am running the following config on Databricks
Scala: 2.12
Spark: 3.1.2

I installed the following jars for mmlspark in Databricks:

- mmlspark-1.0.0-rc4.jar
- mmlspark-core-1.0.0-rc4.jar
- mmlspark-lightgbm-1.0.0-rc4.jar

When I run the following lines, I get a 'Answer from Java side is empty' error:

`from mmlspark.lightgbm import LightGBMClassifier`

`classifier = LightGBMClassifier(
featuresCol="features",
labelCol="ARR_DEL15",
learningRate=0.3,
numIterations=150,
numLeaves=100)`

Are the Jars I have the right ones for spark 3 + ? or is there any other solution to this?

AB#1884836

Start with the Databricks configuration, the listed mmlspark 1.0.0-rc4 JARs, and the Python LightGBMClassifier import and construction. Check compatibility with Scala 2.12 and Spark 3.1.2 and reproduce the “Answer from Java side is empty” failure; done means the cause and a verified resolution are identified.

Written by the indexing model from the issue text.