hive 如何提取json 數組的某一字段

背景：

提取json中某個字段的故事，get_json_object，但是發現好像不能直接提取數組

嘗試使用

split(regexp_replace(substr(contacts, 2, length(contacts)-2), '"}","', '"}",,,,"'), ',,,,')) 

之後提取失敗～ 提取內容基本如下

[{"userId":89224,"contactName":"","contactPhone":"+","originPhone":"+"},

{"userId":89223,"contactName":"","contactPhone":"+","originPhone":""},

{"userId":89221,"contactName":"","contactPhone":"+","originPhone":"+"}]

還是直接放方案吧 '$[*]' 推薦

select user_id, get_json_object(contacts,'$[*].userId') as cons
from XXXX 

你就會得到 類似如下的內容

	user_id	cons
1	107521212	[8397213,85,7418687,304388953,198708241,148,111920,5285539,97773639]

 

但是還不夠 我想要生成映射map

此時推薦 ：

LATERAL VIEW explode (split(substr(t.cons, 2, length(t.cons)-2) , ','))  temp as contact_id 

關於 explode 和  LATERAL VIEW 的用法 可以自行百度 不容踩坑

最終寫法：完成列轉行和映射關係 

SELECT t.user_id as user_id , contact_id from (
select user_id, get_json_object(contacts,'$[*].userId') as cons
from phone_contacts
WHERE dt = '2020-06-27') as t 
LATERAL VIEW explode (split(substr(t.cons, 2, length(t.cons)-2) , ',')) temp as contact_id