I have unit tests using spark local session like so:

SparkSession spark = TestHiveSparkSession.builder().master("local").enableHiveSupport().getOrCreate();

so far we were using it like:

spark.sql("select * from blabla");

However with new Hadoop version it’s no longer an option, we are to migrate to HiveWarehouseSession(Impl).
That session is usually generated as follows:

HiveWarehouseSessionImpl impl = HiveWarehouseSession.session(spark).build();

or

HiveWarehouseSessionImpl impl = HiveWarehouseBuilder.session(spark).build();

but that throws an error saying:

NoSuchElement spark.sql.hive.hiveserver2.jdbc.url

if I put that in the config then there is error saying there’s no driver for my input.

When I put the driver it can’t connect to it.
Which is rightfully so, because I have not started any hive server to connect to.

Is there a way to make it work with local session like it used to, or do I have to redesign the whole thing ?

I am trying to run sqls in unit tests the same way but from the new spark+hive wrapper solution instead of plain spark.

New version is spark-core_2.11, version 2.4.7.7.1.7.2000-305
Old version is spark-core_2.11 version 2.4.0-cdh6.3.4
hadoop went from 3.0.0-cdh6.3.4 to 3.1.1.7.1.7.2000-305