hive解析字符串—get_json_object
简要介绍:在Hive里面使用get_json_object函数来解析json格式数据的具体用法。
语法:
get_json_object(json数据,’$’)
解释:
第一个参数填写json对象变量
第二个参数使用$表示json变量标识
然后用 . 或 [] 读取对象或数组;
如果输入的json字符串无效,那么返回NULL。
每次只能返回一个数据项。
例子:
(0)用到的json字符串介绍:
数组:
[{"name":"巍","age":"20"},{"name":"lan","age":"23"}]
对象:
{
"mylove":
{
"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],
"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]
},
"name":"zhenhun"
}
(一)读取数组内容: $[下标]
hive (g)> select get_json_object('[{"name":"巍","age":"20"},{"name":"lan","age":"23"}]','$[0]');
OK
_c0
{"name":"巍","age":"20"}
(二)读取对象: $.
(1)获取单层值:第二个参数:’$.mylove’
hive (g)> select get_json_object('{"mylove":{"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]},"name":"zhenhun"}','$.mylove');
OK
_c0
{"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]}
Time taken: 0.622 seconds, Fetched: 1 row(s)
hive (g)>
(2)获取多层值:第二个参数:’$.mylove.weilan’
hive (g)> select get_json_object('{"mylove":{"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]},:"zhenhun"}','$.mylove.weilan');
OK
_c0
[{"name":"巍","age":"20"},{"name":"澜","age":"23"}]
Time taken: 0.586 seconds, Fetched: 1 row(s)
(3)获取数组值:第二个参数:’$.mylove.weilan[0]’
hive (g)> select get_json_object('{"mylove":{"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]},"name":"zhenhun"}','$.mylove.weilan[0]');
OK
_c0
{"name":"巍","age":"20"}
Time taken: 0.185 seconds, Fetched: 1 row(s)
hive (g)>
(4)注意:如果第二个参数,有汉字,会出现问题,结果为NULL
hive (g)> select get_json_object('{"mylove":{"weilan":[{"name":"巍","age":"20"},{"name":"澜","age":"23"}],"柯兰":[{"name":"柯","age":"20"},{"name":"兰","age":"23"}]},:"zhenhun"}','$.mylove.柯兰');
OK
_c0
NULL
Time taken: 0.607 seconds, Fetched: 1 row(s)
hive (g)>
PS:如果第二个参数没有中文,出现解析结果如下
OK
_c0
NULL
Time taken: 0.162 seconds, Fetched: 1 row(s)
说明:输入的json字符串无效,所以返回NULL,一定一定检查有没有中文状态输入的