@activescott Yes, we did. We found out that in our application case some data types cause the problem. In addition the error occurs while reading data in spark - e.g. after transformation, but not always at the same point in your code. The cause is how you write your data. We overcome the problem as follows:

1. Checking to which data types our base data will be casted automatically during the "loading to spark" step. For instance in our case spark had problems to read the sql integer type?!

2. Take care how the spark functions or your udf's change the data types. For instance we ran into the error because precision or scale of our decimals changed.