问题是，从spark shell中的hivecontext中，横向视图爆炸不起作用。下面是示例表和示例Spark代码。从Spark帧“vasotherdf”的预期输出是6，但它给出了8。
配置单元表：

CREATE EXTERNAL TABLE `aa`(
`col1` string)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY ','
LINES TERMINATED BY '\n'
STORED AS INPUTFORMAT
'org.apache.hadoop.mapred.TextInputFormat'
OUTPUTFORMAT
'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
LOCATION
'hdfs://nn1001.dev:8020/tmp/aa'

样本数据：

aaa|qq|ww|dd
aaa
aaa|bbb
ccc

配置单元输出：

select count(distinct vother) as vothers from aa LATERAL VIEW explode(split(col1,'\\|')) a as vother;
6

select distinct vother as vothers from rafm.aa LATERAL VIEW explode(split(col1,'\\|')) a as vother;

aaa
bbb
ccc
dd
qq
ww

Spark输出：

val vasOtherDF = hiveContext.sql("select count(distinct vother) as vothers from aa LATERAL VIEW explode(split(col1,'\\|')) a as vother")
output: 8

select distinct vother as vothers from rafm.aa LATERAL VIEW explode(split(col1,'\\|')) a as vother;
aaa
bbb
ccc
dd
qq
ww

val vasOtherDF = hiveContext.sql("select distinct vother as vothers from aa LATERAL VIEW explode(split(col1,'\\|')) a as vother")
scala> vasOtherDF.show
+-------+
|vothers|
+-------+
|      a|
|      b|
|      c|
|      d|
|      q|
|      w|
|      ||
|       |
+-------+