Question

我正在尝试在SQL Server中创建一个使用Posexplode函数在Hive中完成的进程。

我有一个大型数据集，如下所示：

    userid     sku                         qty
1   abc        2427022                      3
2   abc        1883575|2427022             1|3
3   def        2427022|1562336|153842      1|1|1
4   ghi        2427022                       3

我希望数据看起来像：

    userid  sku qty
1   abc 2427022 3
2   abc 1883575 1
3   abc 2427022 3
4   def 2427022 1
5   def 1562336 1
6   def 153842  1
7   ghi 2427022 3

谢谢！

Answer 1

我明白了。如果有人在分隔2个分隔列时出现类似的问题，我使用下面的代码来获得所需的输出。我使用SQL Server中的Cross Apply函数和XML来获得与在Hive中使用posexplode函数相同的输出。

--Create test table
create table Table1 (userid varchar(max), Sku varchar(max), Qty varchar(max))
insert Table1 select 'abc', '2427022' , '3'
insert Table1 select 'abc', '1883575|2427022', '1|3'
insert Table1 select 'def', '2427022|1562336|153842' , '1|1|1'
insert Table1 select 'ghi', '2427022' , ' 3'


;WITH CTE1
AS
(
  SELECT RN= Row_Number () over(Order by userid),userid,  
     Split.a.value('.', 'VARCHAR(100)') AS Sku
 FROM  
 (
     SELECT Userid,Sku,  
         CAST ('<M>' + REPLACE(sku, '|', '</M><M>') + '</M>' AS XML) AS Data  
     FROM  Table1
 ) AS A CROSS APPLY Data.nodes ('/M') AS Split(a)
 )
,CTE2 
AS
(
   SELECT RN= Row_Number () over(Order by userid),userid,  
     Split.b.value('.', 'VARCHAR(100)') AS qty
 FROM  
 (
     SELECT Userid,qty,  
         CAST ('<M>' + REPLACE(qty, '|', '</M><M>') + '</M>' AS XML) AS Data  
     FROM  Table1
 ) AS B CROSS APPLY Data.nodes ('/M') AS Split(b)
 )
 Select c.userid, c.Sku, d.qty
 from CTE1 as c
 Inner Join CTE2 as d
 on c.RN =d.RN

SQL Server功能类似于Hive中的Posexplode？

1 个答案: