HyukjinKwon commented on a change in pull request #28648:
URL: https://github.com/apache/spark/pull/28648#discussion_r431875923
##########
File path: python/pyspark/context.py
##########
@@ -864,8 +865,21 @@ def union(self, rdds):
first_jrdd_deserializer = rdds[0]._jrdd_deserializer
if any(x._jrdd_deserializer != first_jrdd_deserializer for x in rdds):
rdds = [x._reserialize() for x in rdds]
- cls = SparkContext._jvm.org.apache.spark.api.java.JavaRDD
- jrdds = SparkContext._gateway.new_array(cls, len(rdds))
+ gw = SparkContext._gateway
+ jvm = SparkContext._jvm
+ jrdd_cls = jvm.org.apache.spark.api.java.JavaRDD
+ jpair_rdd_cls = jvm.org.apache.spark.api.java.JavaPairRDD
+ jdouble_rdd_cls = jvm.org.apache.spark.api.java.JavaDoubleRDD
+ if is_instance_of(gw, rdds[0]._jrdd, jrdd_cls):
+ cls = jrdd_cls
+ elif is_instance_of(gw, rdds[0]._jrdd, jpair_rdd_cls):
Review comment:
It will still work:
```python
>>> rdd1 = sc.parallelize([1,2,3,4,5])
>>> rdd2 = sc.parallelize([6,7,8,9,10])
>>> pairRDD1 = rdd1.zip(rdd2)
>>> sc.union([rdd1, pairRDD1]).collect()
[1, 2, 3, 4, 5, (1, 6), (2, 7), (3, 8), (4, 9), (5, 10)]
>>> sc.union([pairRDD1, rdd1]).collect()
[(1, 6), (2, 7), (3, 8), (4, 9), (5, 10), 1, 2, 3, 4, 5]
```
because it converts all to `JavaRDD` via `self._reserialize()` when they are
different types.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
