In Spark Streaming, is there a way to write output to different paths based on the partition key? The saveAsTextFiles method will write output in the same directory.
For example, if the partition key has a hour/day column and I want to separate DStream output into different directories by hour/day. -- Chen Song
