如何从配置单元表中的json字符串中提取数组元素？

Posted 2023-04-13

【中文标题】如何从配置单元表中的json字符串中提取数组元素？

【英文标题】：How to extract array elements from json string in hive table?

【发布时间】：2021-06-28 17:23:19

【问题描述】：

我有一个包含 json 字符串的配置单元表：

column
"id": 123, "ShortName":"test tax","LongName":"test","Description":"test","aFactor":"true","bFactor":"true","cFactor":"false"
"id": 456, "ShortName":"test tax","LongName":"test","Description":"test","aFactor":"false","bFactor":"true","cFactor":"true"
"id": 678, "ShortName":"test tax","LongName":"test","Description":"test","aFactor":"false","cFactor":"true"

我想从这里创建一个数组，只要 aFactor、bFactor 或 cFactor 为真：

id	factors
123	aFactor,bFactor
456	bFactor,cFactor
678	cFactor

为 false 或 null 是等价的。

【参考方案1】：

您可以使用以下步骤执行此操作 -

首先确定 *Factor 字段的位置。 如果它们为真则使用，然后将它们连接起来 整个查询应如下所示 -

select get_json_object(str,'$.id') as id,
if(get_json_object(str,'$.aFactor')='true','aFactor','') ||if(get_json_object(str,'$.bFactor')='true',',bFactor','')||if(get_json_object(str,'$.cFactor')='true',',cFactor','') factors
from mytable

示例结果在屏幕截图中。 测试查询

with tab as ( select '"id": 123, "ShortName":"test tax","LongName":"test","Description":"test","aFactor":"true","bFactor":"true","cFactor":"false"' str)
select get_json_object(str,'$.id') as id,
if(get_json_object(str,'$.aFactor')='true','aFactor','') ||  
if(get_json_object(str,'$.bFactor')='true',',bFactor','') ||
if(get_json_object(str,'$.cFactor')='true',',cFactor','') 
 factors