根据JSON数据创建配置单元表 [英] Create hive table from JSON data

问题描述

我有一个包含Json数据的文件，其格式如下:

I have a file with Json data which takes the below form:

例如:

 {
    "Name": "xxxx",
    "Address": [{
        "Street": "aa",
        "City": "bbb"
    }, {
        "Street": "ccc",
        "City": "ddd",
        "Country": "eee"
    }]
}

以上Json是有效的Json.我想使用JsonSerde在上述形式的数据之上创建一个配置单元表.

The above Json is a valid Json. I want to create a hive table on top of data of above form using JsonSerde.

推荐答案

使用已定义的所有可能字段创建表.如果json中不存在该字段，则select将返回NULL:

Create table with all possible fields defined. If field is not present in json, select will return NULL:

CREATE EXTERNAL TABLE your_table (
Name string,
Address array<struct<Street:string,City:string,Country:string>>
)
ROW FORMAT SERDE 'org.openx.data.jsonserde.JsonSerDe'

如果JSON文件中的名称与Hive保留字冲突，则添加映射并在表定义中重命名名称:

if you have names in JSON file which conflict with Hive reserved words then add mapping and rename names in table definition:

WITH SERDEPROPERTIES ('mapping.renamed_column'='original_column')并重命名表列.

将文件放在表格位置.

另请参见带有一些示例的文档: https://github.com/rcongiu/Hive -JSON-Serde

See also docs with some examples here: https://github.com/rcongiu/Hive-JSON-Serde

这篇关于根据JSON数据创建配置单元表的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！