从类似于 ESV 的元数据驱动模式查询数据

Question

我们有一个类似于 EAV 的数据模型。但是，我们的属性集有限。在我们的数据模型中，我们存储了个人完成的不同活动，并且根据 activity 类型，属性会有所不同。

我们想要获取个人的活动和相应的 attribute/values。我已经描述了 Activity、ActivityDetails table 和预期的结果集。

能否请您指导我如何从数据模型中得出预期的结果集？我们必须做一些转变。但是，不确定是否达到相同。

在下面的示例数据中，ClickURL 是一个 Activity。它具有三个属性：URLAddress、ClickCount、ClickDate。三个属性对应的值为：www.companyurl.com,10,20140101

结果集应该如下所示：

ActivityName | Description            | URLAddress          | ClickCount    | ClickDate
-------------------------------------------------------------------------------------------
ClickURL     | Click advertisementURL | www.companyurl.com  |10             |20140101

Our Data model & Expected Result

数据查询示例

    CREATE TABLE [dbo].[Activity](
    [ActivityTypeID] [int] IDENTITY(1,1) NOT NULL,
    [ActivityName] VARCHAR(50) NULL,
    [Description] VARCHAR(255) NULL,
    [StringField1] [varchar](50) NULL,
    [StringField2] [varchar](50) NULL,
    [StringField3] [varchar](50) NULL,  
    [NumField1] [varchar](50) NULL,
    [NumField2] [varchar](50) NULL, 
    [DTTMField1] [varchar](50) NULL,
    [DTTMField2] [varchar](50) NULL
) ON [PRIMARY]
GO

CREATE TABLE [dbo].[ActivityDetails](
    [ActivityDetailID] [int] IDENTITY(1,1) NOT NULL,
    [ActivityTypeID] VARCHAR(50) NULL,
    [StringField1] [varchar](50) NULL,
    [StringField2] [varchar](50) NULL,
    [StringField3] [varchar](50) NULL,  
    [NumField1] [] NULL,
    [NumField2] [int] NULL, 
    [DTTMField1] [datetime] NULL,
    [DTTMField2] [datetime] NULL
) ON [PRIMARY]
GO

INSERT INTO [dbo].[Activity](
[ActivityName]
,[Description] 
,[StringField1]
,[StringField2]
,[StringField3]
,[NumField1] 
,[NumField2] 
,[DTTMField1]
,[DTTMField2]
)
VALUES
('ClickURL','Click advertisementURL','URLAddress',NULL,NULL,'ClickCount',NULL,'clickDate',NULL)
GO

INSERT INTO [dbo].[ActivityDetails](
[ActivityTypeID]
,[StringField1]
,[StringField2]
,[StringField3]
,[NumField1] 
,[NumField2] 
,[DTTMField1]
,[DTTMField2]
)
VALUES
(1,'www.companyurl.com',NULL,NULL,10,NULL,'20140101',NULL)
GO

Answer 1

我必须承认我怀疑你的方法是最好的...但是 - 至少 - 你可以试试这个：

注意： 出于测试目的，我在最后删除了创建的 table。小心真实数据！

--你的table

 CREATE TABLE [dbo].[Activity](
    [ActivityTypeID] [int] IDENTITY(1,1) NOT NULL,
    [ActivityName] VARCHAR(50) NULL,
    [Description] VARCHAR(255) NULL,
    [StringField1] [varchar](50) NULL,
    [StringField2] [varchar](50) NULL,
    [StringField3] [varchar](50) NULL,  
    [NumField1] [varchar](50) NULL,
    [NumField2] [varchar](50) NULL, 
    [DTTMField1] [varchar](50) NULL,
    [DTTMField2] [varchar](50) NULL
) ON [PRIMARY]
GO

CREATE TABLE [dbo].[ActivityDetails](
    [ActivityDetailID] [int] IDENTITY(1,1) NOT NULL,
    [ActivityTypeID] VARCHAR(50) NULL,
    [StringField1] [varchar](50) NULL,
    [StringField2] [varchar](50) NULL,
    [StringField3] [varchar](50) NULL,  
    [NumField1] [int] NULL,
    [NumField2] [int] NULL, 
    [DTTMField1] [datetime] NULL,
    [DTTMField2] [datetime] NULL
) ON [PRIMARY]
GO

INSERT INTO [dbo].[Activity](
[ActivityName]
,[Description] 
,[StringField1]
,[StringField2]
,[StringField3]
,[NumField1] 
,[NumField2] 
,[DTTMField1]
,[DTTMField2]
)
VALUES
('ClickURL','Click advertisementURL','URLAddress',NULL,NULL,'ClickCount',NULL,'clickDate',NULL)
GO

INSERT INTO [dbo].[ActivityDetails](
[ActivityTypeID]
,[StringField1]
,[StringField2]
,[StringField3]
,[NumField1] 
,[NumField2] 
,[DTTMField1]
,[DTTMField2]
)
VALUES
(1,'www.companyurl.com',NULL,NULL,10,NULL,'20140101',NULL)
GO

--我声明一个变量并动态创建语句

DECLARE @cmd VARCHAR(MAX)=
(
    SELECT
    'SELECT ''' + a.ActivityName + ''' AS ActivityName'
    +   ',''' + a.Description + ''' AS Description'
    +   ISNULL(',''' + ad.StringField1 +''' AS ' + QUOTENAME(a.StringField1),'')
    +   ISNULL(',''' + ad.StringField2 +''' AS ' + QUOTENAME(a.StringField2),'')
    +   ISNULL(',''' + ad.StringField3 +''' AS ' + QUOTENAME(a.StringField3),'')
    +   ISNULL(',' + CAST(ad.NumField1 AS VARCHAR(100)) +' AS ' + QUOTENAME(a.NumField1),'')
    +   ISNULL(',' + CAST(ad.NumField2 AS VARCHAR(100)) +' AS ' + QUOTENAME(a.NumField2),'')
    +   ISNULL(',CAST(''' + CONVERT(VARCHAR(100),ad.DTTMField1,126) +''' AS DATETIME) AS ' + QUOTENAME(a.DTTMField1),'')
    +   ISNULL(',CAST(''' + CONVERT(VARCHAR(100),ad.DTTMField2,126) +''' AS DATETIME) AS ' + QUOTENAME(a.DTTMField1),'')
    FROM Activity AS a
    INNER JOIN ActivityDetails AS ad ON a.ActivityTypeID=ad.ActivityTypeID
)

--这里执行语句

EXEC(@cmd)
GO

--清理（注意真实数据！）

--DROP TABLE ActivityDetails;
--DROP TABLE Activity;

创建的语句是这样的：

SELECT 'ClickURL' AS ActivityName
      ,'Click advertisementURL' AS Description
      ,'www.companyurl.com' AS [URLAddress]
      ,10 AS [ClickCount]
      ,CAST('2014-01-01T00:00:00' AS DATETIME) AS [clickDate]

诀窍在于，如果其中一个元素是 NULL，那么一个连接字符串在整个过程中都是 NULL。这就是为什么创建的语句将仅包含使用的列的原因。

Answer 2

谢谢 Shnugo.I 也尝试了下面类似的方法。

DECLARE @ActivityTypeID int =1
DECLARE @ColList VARCHAR(MAX)
DECLARE @ActivityDetailsQuery VARCHAR(MAX)
SELECT @ColList = ''''+ ActivityName + ''' as ActivityName ,''' + Description + ''' as Description'
        +CASE when StringField1 is not null then ',StringField1 AS '+ stringField1 else '' end
       + CASE when StringField2 is not null then ',StringField2 AS '+ stringField2 else '' end
       + CASE when StringField3 is not null then ',StringField3 AS '+ StringField2 else '' end
       + CASE when NumField1 IS NOT NULL then ',NumField1 AS '+ NumField1 else '' end
       + CASE when NumField2 IS NOT NULL then ',NumField2 AS '+ NumField2 else '' end
       + CASE when DttmField1 IS NOT NULL then ',DttmField1 AS '+ DttmField1 else '' end
       + CASE when DttmField2 IS NOT NULL then ',DttmField2 AS '+ DttmField2 else '' end
from dbo.Activity
where ActivityTypeID = @ActivityTypeID  

SET @ActivityDetailsQuery = 'SELECT '+ @ColList + ' FROM dbo.ActivityDetails WHERE ActivityTypeID = ' + CAST(@ActivityTypeID as VARCHAR(10))
EXEC(@ActivityDetailsQuery)

从类似于 ESV 的元数据驱动模式查询数据

Query data from meta data driven schema similar to EAV

sql

sql-server

pivot

database-metadata

entity-attribute-value