Closed qiulin closed 4 years ago
getExistingFilters
is run on the executors, so there is no driver SparkSession
object available. If you're trying to create a JDBC random output, you'd need to use a JDBC client for this method because Spark's JDBC integration can only be used from the driver.
Out of curiosity, why do you want to create this as a random output?
getExistingFilters
is run on the executors, so there is no driverSparkSession
object available. If you're trying to create a JDBC random output, you'd need to use a JDBC client for this method because Spark's JDBC integration can only be used from the driver.Out of curiosity, why do you want to create this as a random output?
Thanks for your answer. We need to support creating Type 2 SCD dimension table on a RDMS Data Warehouse, like Greenplum.
When try to implement getExistingForFilters of RandomOutput and access Spark Session, got NullPointerException.
Code like this:
Exception log like this: