Spark：数据中有null值时的group结果

数据中有null值时的group结果：

Seq(("a", new BigDecimal(1.2)), ("a", null), (null, new BigDecimal(1.2)), (null, new BigDecimal(1.2)) ).toDF("a", "v")

res20.show
/*
+----+--------------------+
|   a|                   v|
+----+--------------------+
|   a|1.199999999999999956|
|   a|                null|
|null|1.199999999999999956|
|null|1.199999999999999956|
+----+--------------------+
*/

res20.groupBy("a").sum("v").show
/*
+----+--------------------+
|   a|              sum(v)|
+----+--------------------+
|null|2.399999999999999912|
|   a|1.199999999999999956|
+----+--------------------+
*/