真的需要ORM框架吗?

在长期的Coding的实践过程中发现未必需要一个ORM框架来作为低层数据访问的工具,实际的业务逻辑所需要的对数据访问的功能往往只是ORM框架一个小的子集,甚至连Mirco级别(如PetaPoco, Dapper)的ORM框架都可以不用。当然这是只针对在一种类型数据库平台上开发而言,不涉及到系统要有跨数据库平台的特性。其实现实系统中真正需要跨数据库平台的系统并不多。

本人并不否认大型的ORM框架如EF、NHibernate对业界所作出的贡献,但在使用之前需要搞清楚:

1. 需要解决的问题是什么?

2. 框架的引入需要付出什么?

3. 有没有更简单的替代方法?

其实我关注的是第3点,在实现项目中我使用过Mirco ORM如Dapper,但逐渐也没再用了,还是直接写数据访问的代码,为了提高编码的效率,编写了一个在SQL Server中使用的存储过程,用来生成C#代码。

 

CREATE PROCEDURE [dbo].[GenerateEntity]
  @tableName VARCHAR(60),
  @getset BIT = 0
AS
BEGIN
    DECLARE @output AS VARCHAR(MAX);
    DECLARE @newline AS VARCHAR(2) =  CHAR(13) + CHAR(10)
    DECLARE @lines AS TABLE ( line VARCHAR(255) )
    DECLARE @line VARCHAR(255)
    
    SET NOCOUNT ON
    
    INSERT INTO @lines
    SELECT 'public ' +
        CASE WHEN DATA_TYPE='varchar' OR DATA_TYPE='nvarchar' THEN 'string'
             WHEN DATA_TYPE='tinyint' THEN 'byte'  
             WHEN DATA_TYPE='bit' THEN 'bool' 
             WHEN DATA_TYPE='datetime' OR DATA_TYPE='datetime2' THEN 'DateTime' 
             ELSE DATA_TYPE END + 
        CASE WHEN IS_NULLABLE='YES' AND NOT (DATA_TYPE='varchar' OR DATA_TYPE='nvarchar') THEN '? ' ELSE ' ' END + COLUMN_NAME +
        CASE WHEN @getset = 0 THEN ';' ELSE ' { get; set; }' END
    FROM  INFORMATION_SCHEMA.COLUMNS WHERE TABLE_NAME = @tableName
    
    SET @output = 'public class ' + @tableName + @newline + '{' + @newline;
    
    DECLARE cur CURSOR FOR SELECT line FROM @lines
    OPEN cur 
    FETCH cur INTO @line
    WHILE @@FETCH_STATUS = 0
    BEGIN
        SET @output += '    ' + @line + @newline
        FETCH cur INTO @line    
    END
    CLOSE cur
    DEALLOCATE cur
    
    --SELECT @output + '}'  -- 在text output 窗口不能完整输出
    SET @output += '}' + @newline 
    
    DECLARE @variableName VARCHAR(255) = LOWER(LEFT(@tableName, 1)) + RIGHT(@tableName, LEN(@tableName)-1)
    
    DELETE @lines
    INSERT INTO @lines
    SELECT @variableName + '.' + COLUMN_NAME + ' = '+
        CASE WHEN IS_NULLABLE='YES' THEN 'reader["'+COLUMN_NAME+'"] == DBNull.Value ? null : ' ELSE '' END +'(' + 
        CASE WHEN DATA_TYPE='varchar' OR DATA_TYPE='nvarchar' THEN 'string'
             WHEN DATA_TYPE='tinyint' THEN 'byte'  
             WHEN DATA_TYPE='bit' THEN 'bool' 
             WHEN DATA_TYPE='datetime' OR DATA_TYPE='datetime2' THEN 'DateTime' 
             ELSE DATA_TYPE END + 
       CASE WHEN IS_NULLABLE='YES' AND NOT (DATA_TYPE='varchar' OR DATA_TYPE='nvarchar') THEN '?' ELSE '' END + ')reader["' + COLUMN_NAME + '"];'
    FROM INFORMATION_SCHEMA.[COLUMNS] WHERE TABLE_NAME = @tableName
    
    SET @output += @tableName + ' ' + @variableName + ' = new ' + @tableName + '();' + @newline;

    DECLARE cur CURSOR FOR SELECT line FROM @lines
    OPEN cur 
    FETCH cur INTO @line
    WHILE @@FETCH_STATUS = 0
    BEGIN
        SET @output += @line + @newline
        FETCH cur INTO @line    
    END
    CLOSE cur
    DEALLOCATE cur
        
    PRINT @output;
END
View Code

如果觉得在SQL生成代码不够爽,可以考虑使用:Razor Engine

http://www.cnblogs.com/artech/p/razor-code-generation.html

http://razorengine.codeplex.com/

posted @ 2013-10-12 09:44  道云  阅读(4319)  评论(1编辑  收藏  举报