Hadoop 配置单元查询嵌套数组
我有一个记录购物车信息的字段,如下所示。数组的长度不是固定的Hadoop 配置单元查询嵌套数组,hadoop,hive,etl,Hadoop,Hive,Etl,我有一个记录购物车信息的字段,如下所示。数组的长度不是固定的 [["SKU_001","Product Name A","Category1",$10.5,1], ["SKU002","Product Name B","Category1",$20,3]] 我的目标是将其反规范化为一个平行,可能如下所示: SKU | Product Name | Category | Unit Price | Product Quantity SKU001 | Product Name A |
[["SKU_001","Product Name A","Category1",$10.5,1],
["SKU002","Product Name B","Category1",$20,3]]
我的目标是将其反规范化为一个平行,可能如下所示:
SKU | Product Name | Category | Unit Price | Product Quantity
SKU001 | Product Name A | Category1 | $10.5 | 1
SKU002 | Product Name B | Category2 | $20 | 3
是否可以在不编写UDF的情况下实现转换?您可以使用“横向视图分解(…)”语法将子数组作为行,然后选择子数组的各个元素作为列
select
cart[0] as SKU,
cart[1] as product_name,
cart[2] as category,
cart[3] as unit_price,
cart[4] as product_quantity
from my_table lateral view explode(my_nested_array_field) t as cart
谢谢你的帮助。我尝试分解嵌套的_数组字段,但执行配置单元查询时出现未知异常错误。我不确定配置单元是否支持嵌套数组查询。我只能使用非嵌套数组(如数组(1,2,3))成功执行分解。还有什么建议吗?我想有些人正在解决这个问题。见此帖: