今天使用EF根据数据库生成模型后，发现之前编写的存储过程并没有返回一个实体类，研究后发现写成了
ALTER proc [dbo].[usp_tree]
@id varchar(max)
AS
declare @str nvarchar(1000);
set @str='with my1 as (select FId,Parent,Name,Code as TagName,leaf from FactoryTree where Parent=('''+@id+''') union all select FactoryTree.FId,FactoryTree.Parent,FactoryTree.Name,FactoryTree.Code as TagName,FactoryTree.leaf from my1,FactoryTree where my1.FId=FactoryTree.Parent and my1.leaf=0 and FactoryTree.leaf=0) select * from my1'
print @str
exec (@str)
这样的写法在EF中默认返回INT类型，而我需要获取多表查询的结果返回一个新的实体类，对其进行操作，所以有新写了一个存储过程
SET ANSI_NULLS ON GO SET QUOTED_IDENTIFIER ON GO ALTER proc [dbo].[EF_Tree] @id varchar(max) AS BEGIN
with my1 as (select FId,Parent,Name,Code as TagName,leaf from FactoryTree where Parent=(''+@id+'') union all select FactoryTree.FId,FactoryTree.Parent,FactoryTree.Name,FactoryTree.Code as TagName,FactoryTree.leaf from my1,FactoryTree where my1.FId=FactoryTree.Parent and my1.leaf=0 and FactoryTree.leaf=0) select * from my1
END
只是去掉了@str就能返回一个实体类，之前使用ADO.NET并没有这种情况。
话说EF还真是好用，改了数据库只要更新下模型就可以了，很HAPPY。
EF调用存储过程
var Result = (from p in db.EF_Tree(Fid)
select p).ToList();
因为Result是实体类，而在项目中要用到DataTable转化JSON（没办法，公司项目里都是用的ADO.NET，我也懒得造*）
下一步，就是将Result转换为DataTable
/// <summary>
/// 将集合类转换成DataTable
/// </summary>
/// <param name="list">集合</param>
/// <returns></returns>
public static DataTable MyListToDataTable(IList list)
{
DataTable result = new DataTable();
if (list.Count > 0)
{
PropertyInfo[] propertys = list[0].GetType().GetProperties();
foreach (PropertyInfo pi in propertys)
{
result.Columns.Add(pi.Name);
}
for (int i = 0; i < list.Count; i++)
{
ArrayList tempList = new ArrayList();
foreach (PropertyInfo pi in propertys)
{
object obj = pi.GetValue(list[i], null);
tempList.Add(obj);
}
object[] array = tempList.ToArray();
result.LoadDataRow(array, true);
}
}
return result;
}
不知道为什么网上的例子里还要在result.Columns.Add(pi.Name，pi.propertyType);
就一直报错DataSet不支持System.nullable<>
后来研究了一下删了pi.propertyType就OK了。
转载于:https://www.cnblogs.com/ckblogs/p/3728330.html