hivesql 将数据处理成复杂json

原创已于 2023-06-27 14:58:08 修改 · 818 阅读

2 ·

CC 4.0 BY-SA版权

文章标签：

#json #sql

于 2023-06-27 14:42:29 首次发布

hive 专栏收录该内容

28 篇文章

订阅专栏

文章展示了如何使用HiveSQL将两种不同类型的数据转换成JSON格式。第一种涉及bankid数组和阶段信息，通过NAMED_STRUCT和COLLECT_LIST构造复杂结构。第二种涉及商品属性，如下装款式和穿搭风格，利用CONCAT和GROUPBY进行组合。

类型一

原数据：bankid是array类型
在这里插入图片描述
目标数据：

{
	"bankname": ["SPDB", "WS_HBBANK", "mytest"],
	"grid": [{
		"name": "阶段1",
		"values": ["38.0,1.0,1.0"]
	}, {
		"name": "阶段2",
		"values": ["11.0,0.0,0.0"]
	}, {
		"name": "阶段3",
		"values": ["11.0,0.0,0.0"]
	}]
}

代码：

SET odps.sql.hive.compatible = true
;

SELECT  
        TO_JSON(NAMED_STRUCT('bankname',bankid,'grid',grid)) AS json
FROM    (
            SELECT  bankid
                    ,COLLECT_LIST(grid) AS grid
            FROM    (
                        SELECT  bankid
                                ,grid
                        FROM    (
                                    SELECT  bankid
                                            ,NAMED_STRUCT('name','阶段1','values',COLLECT_LIST(index_001)) AS index_001
                                            ,NAMED_STRUCT('name','阶段2','values',COLLECT_LIST(index_002)) AS index_002
                                            ,NAMED_STRUCT('name','阶段3','values',COLLECT_LIST(index_003)) AS index_003
                                    FROM    (
                                                SELECT  ARRAY('SPDB','WS_HBBANK','mytest') AS bankid
                                                        ,'38.0,1.0,1.0' AS index_001
                                                        ,'11.0,0.0,0.0' AS index_002
                                                        ,'11.0,0.0,0.0' AS index_003
                                            ) o1
                                    GROUP BY bankid
                                ) o2
                        LATERAL VIEW EXPLODE(ARRAY(index_001,index_002,index_003)) tmp AS grid
                    ) o3
            GROUP BY bankid
        ) o4
;

类型二

原数据：
在这里插入图片描述
目标数据：

{
	"id": 53830,
	"value": [{
		"下装款式": ["直筒裤"]
	}, {
		"下装设计": ["纯色"]
	}, {
		"厚度": ["正常"]
	}, {
		"版型": ["正常"]
	}, {
		"穿搭风格": ["机能", "休闲"]
	}, {
		"裤长": ["长裤"]
	}]
}

代码：

SELECT  CONCAT('{"id":',id,',"value":',result,'}') AS json
FROM    (
            SELECT  id
                    ,CONCAT('[',CONCAT_WS(',',COLLECT_LIST(CONCAT('{"',type,'":',value,'}'))),']') AS result
            FROM    (
                        SELECT  id
                                ,type
                                ,CONCAT('[',CONCAT_WS(',',COLLECT_LIST(CONCAT('"',value,'"'))),']') AS value
                        FROM    temp_wxt_test_0627
                        GROUP BY id
                                 ,type
                    ) o1
            GROUP BY id
        ) o2
;