- 前言
在上一篇《【原创】打造基于Dapper的数据访问层》中,Dapper在应付多表*关联、分组查询、匿名查询等应用场景时经常要手动写SQL语句。看着代码里满屏的红色SQL字符串,简直头大,于是便萌生重复造ORM这个*的念头。本ORM在API设计上最大程度地借鉴 EF 的写法,支持链式查询(点标记)、查询表达式、聚合查询、分组排序、批量插入、批量更新、批量删除、1:1关系外键等。在实体绑定层面,使用 Emit 来动态构建绑定指令,性能最大限度地接近原生水平。
- 性能
7000笔记录循环读1000次,同时加载1:1关系的外键,速度比 EF 稍快。
- 语法
1. 单表查询
// 查询表达式 var query = from a in context.GetTable<Inte_CRM.Demo>() select a; var r1 = query.ToList(); // 点标记 query = context.GetTable<Inte_CRM.Demo>(); r1 = query.ToList(); //SQL=> //SELECT //t0.[DemoId] AS [DemoId], //t0.[DemoCode] AS [DemoCode], //t0.[DemoName] AS [DemoName], //... //t0.[DemoLong] AS [DemoLong], //t0.[DemoLong_Nullable] AS [DemoLong_Nullable] //FROM [Sys_Demo] t0
2. 关联查询
// INNER JOIN var query = from a in context.GetTable<Inte_CRM.CRM_SaleOrder>() join b in context.GetTable<Inte_CRM.Client>() on a.ClientId equals b.ClientId join c in context.GetTable<Inte_CRM.CloudServer>() on b.CloudServerId equals c.CloudServerId select a; var r1 = query.ToList(); // 点标记 query = context .GetTable<Inte_CRM.CRM_SaleOrder>() .Join(context.GetTable<Inte_CRM.Client>(), a => a.ClientId, b => b.ClientId, (a, b) => new { Sale = a, Buyer = b }) .Join(context.GetTable<Inte_CRM.CloudServer>(), b => b.Buyer.CloudServerId, c => c.CloudServerId, (a, c) => new Inte_CRM.CRM_SaleOrder { }) .Where(a => a.ClientId > ); //r1 = query.ToList(); //SQL=> //SELECT //t0.[OrderId] AS [OrderId], //t0.[OrderNo] AS [OrderNo], //t0.[Remark] AS [Remark], //t0.[ClientId] AS [ClientId] //FROM [CRM_SaleOrder] t0 //INNER JOIN [Bas_Client] t1 ON t0.[ClientId] = t1.[ClientId] //INNER JOIN [Sys_CloudServer] t2 ON t1.[CloudServerId] = t2.[CloudServerId] //WHERE t0.[ClientId] > 0
3. 分组分页
// 分组后再分页 query = from a in context.GetTable<Inte_CRM.Client>() where a.ClientName == "TAN" group a by new { a.ClientId, a.ClientName } into g orderby new { g.Key.ClientName, g.Key.ClientId } select new { Id = g.Key.ClientId, Name = g.Min(a => a.ClientId) }; query = query.Skip().Take(); r1 = query.ToList(); //SQL=> //SELECT //t0.[Id], //t0.[Name] //FROM ( // SELECT // t0.[ClientId] AS [Id], // MIN(t0.[ClientId]) AS [Name], // t0.[ClientName] AS [ClientName] // FROM [Bas_Client] t0 // WHERE t0.[ClientName] = N'TAN' // GROUP BY t0.[ClientId],t0.[ClientName] // Having t0.[ClientId] > 0 // ) t0 //ORDER BY t0.[ClientName] //OFFSET 2 ROWS FETCH NEXT 3 ROWS ONLY
4. 批量插入
context.Insert<Inte_CRM.Thin>(collection); context.SubmitChanges(); //SQL=> //INSERT INTO[Sys_Thin] //([ThinId],[ThinName]) //VALUES //(2, N'002'),(3,N'003')
5. 导航属性
// 更简单的赋值方式 // 适用场景:在显示列表时只想显示外键表的一两个字段 query = from a in context.GetTable<Inte_CRM.CRM_SaleOrder>() select new Inte_CRM.CRM_SaleOrder(a) { Client = new Inte_CRM.Client(a.Client) { CloudServer = new Inte_CRM.CloudServer { CloudServerId = a.Client.CloudServer.CloudServerId, CloudServerName = a.Client.CloudServer.CloudServerName } }, HeavyBuyer = new Inte_CRM.Client { ClientId = a.Client.ClientId + , ClientName = a.Client.ClientName + "_heavy", CloudServer = new Inte_CRM.CloudServer { CloudServerId = a.Client.CloudServer.CloudServerId + , CloudServerName = a.Client.CloudServer.CloudServerName + "_heavy", } } }; r1 = query.ToList(); //SQL=> //SELECT //t0.[OrderId] AS [OrderId], //t0.[OrderNo] AS [OrderNo], //t0.[Remark] AS [Remark], //t0.[ClientId] AS [ClientId], //t1.[ClientId] AS [ClientId1], //t1.[ClientCode] AS [ClientCode], //t1.[ClientName] AS [ClientName], //t1.[State] AS [State], //t1.[ActiveDate] AS [ActiveDate], //t1.[CloudServerId] AS [CloudServerId], //t2.[CloudServerId] AS [CloudServerId1], //t2.[CloudServerName] AS [CloudServerName], //t1.[ClientId] + 10 AS [ClientId2], //t1.[ClientName] + N'_heavy' AS [ClientName1], //t2.[CloudServerId] + 10 AS [CloudServerId2], //t2.[CloudServerName] + N'_heavy' AS [CloudServerName1] //FROM [CRM_SaleOrder] t0 //LEFT JOIN [Bas_Client] t1 ON t0.[ClientId] = t1.[ClientId] //LEFT JOIN [Sys_CloudServer] t2 ON t1.[CloudServerId] = t2.[CloudServerId]
其它更多示例在源码的 demo 中有详细说明,源码地址:https://github.com/TANZAME/Inte.XFramework