HiveQL：合并数组列

Question

我有一个表，其中的列具有数组，如何编写HiveQL合并数组列？

| id | colA             | colB            | colC 
+----+------------------+-----------------+------------------    
| 1  | ["john", "james"]| ["peter"]       | ["sam","peter"]
| 2  | ["jane"]         | ["doug"]        | ["mary","peter"]
| 3  | ["jan", "james"] | ["peter","mary"]| ["sam","peter"]

编写查询以显示如下数据：

| id | newcol
+----+------------------------------------------------------    
| 1  | ["john", "james", "peter", "sam","peter"]
| 2  | ["jane", "doug", "mary","peter"]
| 3  | ["jan", "james", "peter","mary","sam","peter"]

Answer 1

如果要从多个字符串数组中创建一个新的字符串数组，可以使用以下内容（请注意，您需要检查colB和colC是否为空，以避免产生多余的逗号）：

split (
concat (

concat_ws(',',colA),
if(size(colB)>0, concat(',', concat_ws(',',colB)), ''    ),
if(size(colC)>0, concat(',', concat_ws(',',colC)), ''    )


),',')

Answer 2

hive> select split(concat_ws(',',array("john", "james"), array('peter'), array("sam","peter")), ",") as a;
OK
["john","james","peter","sam","peter"]

根据您的情况。

select split(concat_ws(',',ColA, ColB, ColC), ",") as a;

Answer 3

类似的东西：

Concat or ||

SELECT id, colA||','||colB||','||colC AS newcol FROM myTable

OR

SELECT id, Concat(colA,',',colB,',',colC,) AS newcol FROM myTable