[Docs] Fix typo for Python section on unifying Kafka streams

[theckang]

1. kafkaStreams is a list. The list should be unpacked when passing it into the streaming context union method, which accepts a variable number of streams.
2. print() should be pprint() for pyspark.

This contribution is my original work, and I license the work to the project under the project's open source license.

[srowen]

LGTM

[SparkQA]

Test build #2010 has finished for PR 9545 at commit 8020f13.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[srowen]

@C-kang actually I don't know Python syntax well enough to be sure about this -- this is varargs syntax?

[theckang]

@srowen right. The method is: def union(self, *dstreams), so it expects a variable number of dstreams passed in as arguments. If you pass in a list of dstreams without unpacking it, then vargs= [kafkaStreams, ] as opposed to vargs = [stream1, stream2, stream3, ...]. In the method then, len(vargs) == 1, so it just returns [kafkaStreams] as the unifiedStream without actually performing the union. And then unifiedStream.pprint() will fail because unifiedStream is not a dstream but a list.

[srowen]

Merged to master/1.6/1.5