hive中解析json常用的方式按效率由低到高有三种:
- regexp_extract
- get_json_object
- json_tuple
demo数据:
set hivevar:person={"name":"amos","sex":"man","age":32};
1.regexp_extract 解析,这种方式很不友好,需要数据有固定的格式
select regexp_extract(regexp_replace('${hivevar:person}','\"',''),'name:(.*),sex',1) as name;
OK
name
amos
2.get_json_object 解析
select get_json_object('${hivevar:person}','$.name') as name,get_json_object('${hivevar:person}','$.sex') as sex,get_json_object('${hivevar:person}','$.age') as age;
OK
name sex age
amos man 32
3.json_tuple解析
with json_table as(select '${hivevar:person}' as person)
select person.* from json_table lateral view json_tuple(person,'name','sex','age') person as name,sex,age;
OK
person.name person.sex person.age
amos man 32