SQL查询转换为JSON格式

Question

您好，我是初学JSON的人，想问一个问题：

我的目标是尝试使结果看起来像这样：

{
    "userid": "10293304" ,     "segments": ["Online_Flag"]
} 

{
    "userid": "10292298" , "segments": ["schedule_Appointment", "Enrolled_Order","Complete_Order"]

}

我可以使用 T-SQL 查询，使结果看起来像这样：

[{"userid":"10293159","segments":[{"segment":"Schedule_Appointment"}]},
{"userid":"10293056","segments":[{"segment":"Schedule_Appointment"}]},
{"userid":"10292838","segments":[{"segment":"Schedule_Appointment"}]},

问题在于我只需要显示JSON值（Schedule_Appointment），而不是键（segment）。

是否有任何方法（使用TSQL）可以隐藏键，仅在JSON数组中显示值？

我已经尝试查找使用SQL创建JSON数组的方法，但没有成功，如果有人能解释一下，我会非常感激。

SQL SERVER 2016

数据

CREATE TABLE [dbo].[Lotame_JSON](
[lead_id] [varchar](100) NULL,
[dist_date_key] [date] NULL,
[online_flag] [int] NULL,
[sched_appt_qty] [int] NULL,
[enroll_order_qty] [int] NULL,
[compl_order_qty] [int] NULL)

insert into [dbo].[Lotame_JSON]
values('105646','2016-12-1',1,0,0,1)

insert into [dbo].[Lotame_JSON]
values('125646','2016-12-1',0,0,1,0)

insert into [dbo].[Lotame_JSON]
values('112646','2016-12-1',0,1,1,0)

insert into [dbo].[Lotame_JSON]
values('106446','2016-12-1',0,0,1,0)

我的问题是：

select 
[lead_id] AS 'userid',
(
SELECT segment
    from 
    (
        SELECT              
            CASE WHEN [online_flag] > 0 THEN '1' else null end as 'online_flag',
            CASE WHEN sched_appt_qty > 0 THEN '2' else null end as 'Schedule_Appointment',
            CASE WHEN enroll_order_qty > 0 THEN '3'  else null end as 'Enrolled_Order',
            CASE WHEN compl_order_qty > 0 THEN '4'  else null end as 'Complete_Order'
        FROM [dbo].[Lotame_JSON] as sub
        WHERE sub.lead_id = main.lead_id
    ) t
    UNPIVOT
    (
        segment1
        for segment in (online_flag,Schedule_Appointment,Enrolled_Order,Complete_Order)
    ) as UnPvot
    for JSON PATH
) AS 'segments' 
from [dbo].[Lotame_JSON] as main
where online_flag = 0 and (sched_appt_qty > 0 or enroll_order_qty > 0   or    compl_order_qty > 0)
and dist_date_key >= '2016-9-1'
FOR JSON PATH

谢谢

- nghoang

---

你正在使用哪个版本的 SQL SERVER？ - Pரதீப்

另外，您可以添加您已使用的查询以获取到目前为止的结果。 - NotMe

更新了，感谢你的帮助！！ - nghoang

在这里，发布代码、数据或查询结果的图片而非文本被认为是非常粗鲁的行为。 - Joel Coehoorn

@JoelCoehoorn 已更新，感谢您的建议，非常感谢。 - nghoang

Answer 1

你已经很接近了。在SQL中正确格式化JSON数组有一个技巧。你需要使用JSON_QUERY、STUFF和FOR XML PATH的组合来处理你想要放入数组中的数据。如果想要得到不带最外层方括号的结果，可以使用FOR JSON PATH, WITHOUT_ARRAY_WRAPPER。将它们结合起来，你就可以得到如下查询：

SELECT 
userid = [lead_id]
,segments = JSON_QUERY(
                        '[' + STUFF(
                                    (
                                        SELECT ',' + '"' +  segment + '"'
                                        FROM
                                        (
                                            SELECT              
                                                CASE WHEN [online_flag] > 0 THEN '1' ELSE NULL END AS 'online_flag',
                                                CASE WHEN sched_appt_qty > 0 THEN '2' ELSE NULL END AS 'Schedule_Appointment',
                                                CASE WHEN enroll_order_qty > 0 THEN '3' ELSE NULL END AS 'Enrolled_Order',
                                                CASE WHEN compl_order_qty > 0 THEN '4' ELSE NULL END AS 'Complete_Order'
                                            FROM [dbo].[Lotame_JSON] AS sub
                                            WHERE sub.lead_id = main.lead_id
                                        ) t
                                        UNPIVOT
                                        (
                                            segment1 FOR segment IN (online_flag, Schedule_Appointment, Enrolled_Order, Complete_Order)
                                        ) AS UnPvot
                                        FOR XML PATH ('')
                                    )
                        , 1, 1, '') + ']'
                    )
FROM [dbo].[Lotame_JSON] AS main
WHERE online_flag = 0
    AND dist_date_key >= '2016-9-1'
    AND (
            sched_appt_qty > 0 
            OR enroll_order_qty > 0
            OR compl_order_qty > 0
        )
FOR JSON PATH, WITHOUT_ARRAY_WRAPPER

- digital.aaron