which matches some of the other helper functions in the API.
Aside from that there are a lot of cases where the encoder<>() function cannot find the encoder for a certain type like Array<DoubleArray> for instance. For these cases I added a fallback on the JavaSparkContext.broadcast function already present in Spark. This makes sure that at the encoder<>() method is attempted (so broadcasting Data classes still works fine), but if that fails it still has a backup solution.
Instead of calling
you can now simply call
which matches some of the other helper functions in the API.
Aside from that there are a lot of cases where the
encoder<>()
function cannot find the encoder for a certain type likeArray<DoubleArray>
for instance. For these cases I added a fallback on theJavaSparkContext.broadcast
function already present in Spark. This makes sure that at theencoder<>()
method is attempted (so broadcasting Data classes still works fine), but if that fails it still has a backup solution.Tests are included of course.