Hive处理Json相关数据

get_json_object 函数

此函数为常用函数，可以取出json数据中的相应的字段的数值

-- get_json_object(json,'$.name') as name
select get_json_object('{"name":"zj","age":25}','$.name') as name ,get_json_object('{"name":"zj","age":25}','$.age') as age;

select get_json_object('{"name":"zj","age":25,"add":{"provinces":"河南","city":"周口"}}','$.add.city');

json_tuple 函数

select json_tuple('{"name":"zj","age":25}','name','age') ;

select b.c0 as name,b.c1 as age from (select json_tuple('{"name":"zj","age":25}','name','age'))b ;

Json 数组

-- 相当于UDTF “一进多出”
SELECT explode(split(regexp_replace(regexp_replace('[{"website":"baidu.com","name":"百度"},{"website":"google.com","name":"谷歌"}]', '\\[|\\]',''),'\\}\\,\\{','\\}\\;\\{'),'\\;'));
-- 说明:  
--     第一次替换：将'['或者']'替换为''
-- 	   第二次替换：将'},{' 替换为'};{'
--     切割：按照';'进行切割

此方式也可以采用编写UDTF来实现，具体步骤可以查看Hive UDF GDF UDTF编写与实现