本文转自:http://www.cnblogs.com/xiepeixing/p/5275999.html

Working with Transactions (EF6 Onwards)

This document will describe using transactions in EF6 including the enhancements we have added since EF5 to make working with transactions easy.

What EF does by default

In all versions of Entity Framework, whenever you execute SaveChanges() to insert, update or delete on the database the framework will wrap that operation in a transaction. This transaction lasts only long enough to execute the operation and then completes. When you execute another such operation a new transaction is started.

Starting with EF6 Database.ExecuteSqlCommand() by default will wrap the command in a transaction if one was not already present. There are overloads of this method that allow you to override this behavior if you wish. Also in EF6 execution of stored procedures included in the model through APIs such as ObjectContext.ExecuteFunction() does the same (except that the default behavior cannot at the moment be overridden).

In either case, the isolation level of the transaction is whatever isolation level the database provider considers its default setting. By default, for instance, on SQL Server this is READ COMMITTED.

Entity Framework does not wrap queries in a transaction.

This default functionality is suitable for a lot of users and if so there is no need to do anything different in EF6; just write the code as you always did.

However some users require greater control over their transactions – this is covered in the following sections.

How the APIs work

Prior to EF6 Entity Framework insisted on opening the database connection itself (it threw an exception if it was passed a connection that was already open). Since a transaction can only be started on an open connection, this meant that the only way a user could wrap several operations into one transaction was either to use a TransactionScope or use the ObjectContext.Connection property and start calling Open() and BeginTransaction() directly on the returned EntityConnection object. In addition, API calls which contacted the database would fail if you had started a transaction on the underlying database connection on your own.

Note: The limitation of only accepting closed connections was removed in Entity Framework 6. For details, see Connection Management (EF6 Onwards).

Starting with EF6 the framework now provides:

  1. Database.BeginTransaction() : An easier method for a user to start and complete transactions themselves within an existing DbContext – allowing several operations to be combined within the same transaction and hence either all committed or all rolled back as one. It also allows the user to more easily specify the isolation level for the transaction.
  2. Database.UseTransaction() : which allows the DbContext to use a transaction which was started outside of the Entity Framework.

Combining several operations into one transaction within the same context

Database.BeginTransaction() has two overrides – one which takes an explicit IsolationLevel and one which takes no arguments and uses the default IsolationLevel from the underlying database provider. Both overrides return a DbContextTransaction object which provides Commit() and Rollback() methods which perform commit and rollback on the underlying store transaction.

The DbContextTransaction is meant to be disposed once it has been committed or rolled back. One easy way to accomplish this is the using(…) {…} syntax which will automatically call Dispose() when the using block completes:

using System;  using System.Collections.Generic;  using System.Data.Entity;  using System.Data.SqlClient;  using System.Linq;  using System.Transactions;    namespace TransactionsExamples  {      class TransactionsExample      {          static void StartOwnTransactionWithinContext()          {              using (var context = new BloggingContext())              {                  using (var dbContextTransaction = context.Database.BeginTransaction())                  {                      try                      {                          context.Database.ExecuteSqlCommand(                              @"UPDATE Blogs SET Rating = 5" +                                  " WHERE Name LIKE '%Entity Framework%'"                              );                            var query = context.Posts.Where(p => p.Blog.Rating >= 5);                          foreach (var post in query)                          {                              post.Title += "[Cool Blog]";                          }                            context.SaveChanges();                            dbContextTransaction.Commit();                      }                      catch (Exception)                      {                          dbContextTransaction.Rollback();                      }                  }              }          }      }  }

Note: Beginning a transaction requires that the underlying store connection is open. So calling Database.BeginTransaction() will open the connection  if it is not already opened. If DbContextTransaction opened the connection then it will close it when Dispose() is called.

Passing an existing transaction to the context

Sometimes you would like a transaction which is even broader in scope and which includes operations on the same database but outside of EF completely. To accomplish this you must open the connection and start the transaction yourself and then tell EF a) to use the already-opened database connection, and b) to use the existing transaction on that connection.

To do this you must define and use a constructor on your context class which inherits from one of the DbContext constructors which take i) an existing connection parameter and ii) the contextOwnsConnection boolean.

Note: The contextOwnsConnection flag must be set to false when called in this scenario. This is important as it informs Entity Framework that it should not close the connection when it is done with it (e.g. see line 4 below):

using (var conn = new SqlConnection("..."))  {      conn.Open();      using (var context = new BloggingContext(conn, contextOwnsConnection: false))      {      }  }

Furthermore, you must start the transaction yourself (including the IsolationLevel if you want to avoid the default setting) and let the Entity Framework know that there is an existing transaction already started on the connection (see line 33 below).

Then you are free to execute database operations either directly on the SqlConnection itself, or on the DbContext. All such operations are executed within one transaction. You take responsibility for committing or rolling back the transaction and for calling Dispose() on it, as well as for closing and disposing the database connection. E.g.:

using System;  using System.Collections.Generic;  using System.Data.Entity;  using System.Data.SqlClient;  using System.Linq;  sing System.Transactions;    namespace TransactionsExamples  {       class TransactionsExample       {          static void UsingExternalTransaction()          {              using (var conn = new SqlConnection("..."))              {                 conn.Open();                   using (var sqlTxn = conn.BeginTransaction(System.Data.IsolationLevel.Snapshot))                 {                     try                     {                         var sqlCommand = new SqlCommand();                         sqlCommand.Connection = conn;                         sqlCommand.Transaction = sqlTxn;                         sqlCommand.CommandText =                             @"UPDATE Blogs SET Rating = 5" +                              " WHERE Name LIKE '%Entity Framework%'";                         sqlCommand.ExecuteNonQuery();                           using (var context =                            new BloggingContext(conn, contextOwnsConnection: false))                          {                              context.Database.UseTransaction(sqlTxn);                                var query =  context.Posts.Where(p => p.Blog.Rating >= 5);                              foreach (var post in query)                              {                                  post.Title += "[Cool Blog]";                              }                             context.SaveChanges();                          }                            sqlTxn.Commit();                      }                      catch (Exception)                      {                          sqlTxn.Rollback();                      }                  }              }          }      }  }

Notes:

  • You can pass null to Database.UseTransaction() to clear Entity Framework’s knowledge of the current transaction. Entity Framework will neither commit nor rollback the existing transaction when you do this, so use with care and only if you’re sure this is what you want to do.
  • You will see an exception from Database.UseTransaction() if you pass a transaction:
    • When the Entity Framework already has an existing transaction
    • When Entity Framework is already operating within a TransactionScope
    • Whose connection object is null (i.e. one which has no connection – usually this is a sign that that transaction has already completed)
    • Whose connection object does not match the Entity Framework’s connection.

Using transactions with other features

This section details how the above transactions interact with:

  • Connection resiliency
  • Asynchronous methods
  • TransactionScope transactions

Connection Resiliency

The new Connection Resiliency feature does not work with user-initiated transactions. For details, see Limitations with Retrying Execution Strategies.

Asynchronous Programming

The approach outlined in the previous sections needs no further options or settings to work with the asynchronous query and save methods. But be aware that, depending on what you do within the asynchronous methods, this may result in long-running transactions – which can in turn cause deadlocks or blocking which is bad for the performance of the overall application.

TransactionScope Transactions

Prior to EF6 the recommended way of providing larger scope transactions was to use a TransactionScope object:

using System.Collections.Generic;  using System.Data.Entity;  using System.Data.SqlClient;  using System.Linq;  using System.Transactions;    namespace TransactionsExamples  {      class TransactionsExample      {          static void UsingTransactionScope()          {              using (var scope = new TransactionScope(TransactionScopeOption.Required))              {                  using (var conn = new SqlConnection("..."))                  {                      conn.Open();                        var sqlCommand = new SqlCommand();                      sqlCommand.Connection = conn;                      sqlCommand.CommandText =                          @"UPDATE Blogs SET Rating = 5" +                              " WHERE Name LIKE '%Entity Framework%'";                      sqlCommand.ExecuteNonQuery();                        using (var context =                          new BloggingContext(conn, contextOwnsConnection: false))                      {                          var query = context.Posts.Where(p => p.Blog.Rating > 5);                          foreach (var post in query)                          {                              post.Title += "[Cool Blog]";                          }                          context.SaveChanges();                      }                  }                    scope.Complete();              }          }      }  }

The SqlConnection and Entity Framework would both use the ambient TransactionScope transaction and hence be committed together.

Starting with .NET 4.5.1 TransactionScope has been updated to also work with asynchronous methods via the use of the TransactionScopeAsyncFlowOption enumeration:

using System.Collections.Generic;  using System.Data.Entity;  using System.Data.SqlClient;  using System.Linq;  using System.Transactions;    namespace TransactionsExamples  {      class TransactionsExample      {          public static void AsyncTransactionScope()          {              using (var scope = new TransactionScope(TransactionScopeAsyncFlowOption.Enabled))              {                  using (var conn = new SqlConnection("..."))                  {                      await conn.OpenAsync();                        var sqlCommand = new SqlCommand();                      sqlCommand.Connection = conn;                      sqlCommand.CommandText =                          @"UPDATE Blogs SET Rating = 5" +                              " WHERE Name LIKE '%Entity Framework%'";                      await sqlCommand.ExecuteNonQueryAsync();                        using (var context = new BloggingContext(conn, contextOwnsConnection: false))                      {                          var query = context.Posts.Where(p => p.Blog.Rating > 5);                          foreach (var post in query)                          {                              post.Title += "[Cool Blog]";                          }                            await context.SaveChangesAsync();                      }                  }              }          }      }  }

There are still some limitations to the TransactionScope approach:

  • Requires .NET 4.5.1 or greater to work with asynchronous methods.
  • It cannot be used in cloud scenarios unless you are sure you have one and only one connection (cloud scenarios do not support distributed transactions).
  • It cannot be combined with the Database.UseTransaction() approach of the previous sections.
  • It will throw exceptions if you issue any DDL (e.g. because of a Database Initializer) and have not enabled distributed transactions through the MSDTC Service.

Advantages of the TransactionScope approach:

  • It will automatically upgrade a local transaction to a distributed transaction if you make more than one connection to a given database or combine a connection to one database with a connection to a different database within the same transaction (note: you must have the MSDTC service configured to allow distributed transactions for this to work).
  • Ease of coding. If you prefer the transaction to be ambient and dealt with implicitly in the background rather than explicitly under you control then the TransactionScope approach may suit you better.

In summary, with the new Database.BeginTransaction() and Database.UseTransaction() APIs above, the TransactionScope approach is no longer necessary for most users. If you do continue to use TransactionScope then be aware of the above limitations. We recommend using the approach outlined in the previous sections instead where possible.

 
 
分类: C#,框架

[转]Working with Transactions (EF6 Onwards)的更多相关文章

  1. Working with Transactions (EF6 Onwards)

    Data Developer Center > Learn > Entity Framework > Get Started > Working with Transactio ...

  2. Testing with a mocking framework (EF6 onwards)

    When writing tests for your application it is often desirable to avoid hitting the database.  Entity ...

  3. Code-Based Configuration (EF6 onwards)

    https://msdn.microsoft.com/en-us/data/jj680699#Using

  4. Entityframework 事务

    Working with Transactions (EF6 Onwards) This document will describe using transactions in EF6 includ ...

  5. Entity Framework 项目使用心得

    在博客园很久了,一直只看不说,这是发布本人的第一个博客. 总结一下在项目中,EntityFramework使用的一下经验拿来和大家分享,希望对大家有用~ 1.         在Entity Fram ...

  6. EF Working with Transactions

    原文:https://msdn.microsoft.com/en-us/data/dn456843.aspx Prior to EF6 Entity Framework insisted on ope ...

  7. EF6 DataMigration 从入门到进阶

    引言 在EntityFramework的开发过程中我们有时因需求变化或者数据结构设计的变化经常会改动表结构.但数据库Schema发生变化时EF会要求我们做DataMigration 和UpdateDa ...

  8. Entity Framework的启动速度优化

    最近开发的服务放到IIS上寄宿之后,遇到一些现象,比如刚部署之后,第一次启动很慢:程序放置一会儿,再次请求也会比较慢.比如第一个问题,可以解释为初次请求某一个服务的时候,需要把程序集加载到内存中可能比 ...

  9. EntityFramework 如何进行异步化(关键词:async·await·SaveChangesAsync·ToListAsync)

    应用程序为什么要异步化?关于这个原因就不多说了,至于现有项目中代码异步化改进,可以参考:实际案例:在现有代码中通过async/await实现并行 这篇博文内容针对的是,EntityFramework ...

随机推荐

  1. Java显式锁学习总结之一:概论

    我们都知道在java中,当多个线程需要并发访问共享资源时需要使用同步,我们经常使用的同步方式就是synchronized关键字,事实上,在jdk1.5之前,只有synchronized一种同步方式.而 ...

  2. 小程序的一个tab切换

    <view class="tab-left" bindtap="tab"> <view class="{{tabArr.curHdI ...

  3. LeetCode解题报告—— Reverse Nodes in k-Group && Sudoku Solver

    1. Reverse Nodes in k-Group Given a linked list, reverse the nodes of a linked list k at a time and ...

  4. Matlab处理数据导出Paraview可读的vtk文件(二)

    由于我在用SPH方法仿真时用的是FORTRAN语言,并且没有找到直接输出vtk文件的代码,因此偷懒通过MATLAB转换一下数据. 用到的Matlab子程序可通过一下链接找到. Matlab处理数据导出 ...

  5. 插入排序(InsertionSort)

    算法描述 插入排序是在一个已经有序的数据序列,要求在这个已经排好的数据序列中插入一个数,但要求插入后此数据序列仍然有序.插入排序是一种稳定的排序. 基本思想 插入排序是在一个已经有序的小序列的基础上, ...

  6. Apache配置文件相关命令

    转:http://www.365mini.com/page/apache-options-directive.htm Options指令是Apache配置文件中一个比较常见也比较重要的指令,Optio ...

  7. npm 安装或更新模块失败的解决办法

    头一次关注npm,在刚刚安装的机子上使用更新指令却报错,我还以为是SHA512有什么问题,后来发现是因为一些网络原因,所以,如果出现错误,务必修改默认配置为国内镜像,但是在publish之前,记得要改 ...

  8. Codeforces Round #274 (Div. 2) Riding in a Lift(DP 前缀和)

    Riding in a Lift time limit per test 2 seconds memory limit per test 256 megabytes input standard in ...

  9. 使用ICSharpCode.SharpZipLib+Aspose模板批量导出Word

    由于是Web端的项目,所以点击按钮之后直接从Aspose模板读取数据,然后在内存中操作,而不是下载到本地后再打包弄到内存中下载.废话不多说,直接上代码 public ActionResult Expo ...

  10. MAC安装go

    下载官方pkg文件,一路傻瓜next就行了. pkg默认是安装到/usr/local/go 安装完成,接下来是配置mac下的环境变量.打开终端(Terminal),敲入一下代码:cd ~ #进入当前用 ...