Question

我的数据库中有几个表格格式为PREFIX_FeatureName。例如：

Processed_CallingCountry（featureName nvarchar（max）null，escalatedCount int null，featureValue nvarchar（max）null，RequestNumber int null）
Processed_CallingCity（featureName nvarchar（max）null，escalatedCount int null，featureValue nvarchar（max）null，RequestNumber int null）

[...]

目前，每个表都有一个存储过程，用于计算该功能，并将其插入到正确的表中。每个功能的处理代码几乎相同，功能名称不同。

示例：

SELECT
    (SELECT SUM(IsSpecial) 
     FROM [dbo].[FeaturesBase] 
     WHERE CreatedTime < a.CreatedTime 
       AND CallingCountry = a.CallingCountry ) AS EscalatedCount,
    'CallingCountry' AS FeatureName,
    RequestNumber,
    CallingCountry AS FeatureValue
FROM 
    [dbo].[CssFeaturesBase] a

此代码针对所有功能进行了复制粘贴，仅更改了功能名称 - 在前面的示例中，＆＃39; CallingCountry＆＃39;将被其他字符串替换，每个存储过程一个。

有没有办法创建一个将字符串作为输入的单个存储过程，并使用字符串/ nvarchar作为参数进行计算？

基本上，我该怎么做：

CREATE PROCEDURE process (@featureName nvarchar(max))
AS 
BEGIN
    SELECT INTO ('Processed' + @featureName)  -- use the table defined by the string
        (SELECT SUM(IsSpecial) 
         FROM [dbo].[FeaturesBase] 
         WHERE CaseCreatedTime < a.CaseCreatedTime 
           AND @featureName = a.@featureName) AS EscalatedCount,  --use @featureName as column name
        @featureName AS FeatureName,   --use the string literal here
        RequestNumber,
        @featureName AS FeatureValue  --select the proper value
    FROM
        [dbo].[FeaturesBase] a

Answer 1

您可能可以使用动态SQL来解决这个问题：

像

这样的东西

CREATE PROCEDURE process (@featureName nvarchar(max))
AS 
BEGIN

    declare @sql nvarchar(max);

    SET @sql = 'SELECT INTO (Processed_'+ @featureName + ') 
        (SELECT SUM(IsSpecial) FROM [dbo].[FeaturesBase] where CaseCreatedTime < a.CaseCreatedTime AND ' + @featureName +' =  a.'+ @featureName+ ')  as EscalatedCount,  
        '''+@featureName+''' AS FeatureName,
        RequestNumber,
        ' + @featureName + ' AS FeatureValue 
    from [dbo].[FeaturesBase] a'

    exec(@sql);
End

我已经测试了这个，它应该执行以下

SELECT INTO (Processed_CallingCountry) 
        (SELECT SUM(IsSpecial) FROM [dbo].[FeaturesBase] where CaseCreatedTime < a.CaseCreatedTime AND CallingCountry =  a.CallingCountry)  as EscalatedCount,  
        'CallingCountry' AS FeatureName,
        RequestNumber,
        CallingCountry AS FeatureValue 
    from [dbo].[FeaturesBase] a

您应该可以在https://msdn.microsoft.com/en-GB/library/ms188001.aspx

找到更多信息

Answer 2

在讨论如何插入数据之前，需要解决一些问题：

为什么featureName为NVARCHAR(MAX)？它会有超过4000个字符吗？我怀疑它应该更接近NVARCHAR(20)。
为什么featureName允许NULL？不应该是NOT NULL？
featureName字段很可能甚至不在这些表格中。您需要一个名为Feature的查找表，其中包含FeatureID INT NOT NULL PRIMARY KEY和FeatureName NVARCHAR(20) NOT NULL。然后，这些其他表只有FeatureID INT字段和定义引用Feature(FeatureID)的外键。
您很可能不需要针对不同功能的单独表格？它们具有完全相同的结构，并且您已经有一个字段（即featureName）来区分行。为什么不只有一个Processed表，其中包含以下字段：(FeatureID INT NOT NULL, escalatedCount int null, featureValue nvarchar(max) null, RequestNumber int null)

SQL - 插入由nvarchar

2 个答案: