I've seen version conflict between Spark's Hadoop dependency and Beeju's Hadoop dependency. (This is actually observed when we change the version of Spark.)
We might have to get rid of Beeju if Spark and Beeju are being conflicted and we are forced to choose one. In other words, to leverage Beeju, tests should run correctly even if we exclude Hadoop dependencies from Beeju.
I've seen version conflict between Spark's Hadoop dependency and Beeju's Hadoop dependency. (This is actually observed when we change the version of Spark.)
We might have to get rid of Beeju if Spark and Beeju are being conflicted and we are forced to choose one. In other words, to leverage Beeju, tests should run correctly even if we exclude Hadoop dependencies from Beeju.