C#使用BerkeleyDB操作简介

Berkeley DB是历史悠久的嵌入式数据库系统，主要应用在UNIX/LINUX操作系统上。Berkeley DB的存储的是key/value键值对，可以理解为硬盘上的超级hash表。其可以管理256TB数据，而且能支撑几千个并发访问。目前Berkeley DB有C++版和Java版。所以，我们需要一个访问的中间转换，已经有人发布了C#的API。可以从 Berkeley DB for .NET 上面找到，现在最新版是0.95版本，可以支持4.3和4.5版。本篇将以4.5版做实例。BerkeleyDB的版本可以在http://www.oracle.com/technology/products/berkeley-db/index.html下载，当前最新版本为4.7版。4.5 C++版的Berkeley DB可以在http://www.oracle.com/technology/software/products/berkeley-db/db/index.html这里下载。

By Birdshover@ 博客园 http://www.cnblogs.com/birdshover/

下载到Berkeley DB for .Net的API——libdb-dotnet_0_95.zip后，就可以开始使用了。首先在libdb-dotnet_0_95.zip解压缩的bin目录找到libdb_dotNET45.dll，这个就是4.5版本使用的dll。新建项目，引用这个dll。注意，自己编译源码可能会编译不过，主要是因为里面一些委托和委托的参数可见性不一致造成的。把那些参数用到的class 或者struct都调成public即可。

BerkeleyDB的数据库操作需要借助DbBTree类。因此需要先得到DbBTree的实例，但是DbBTree类会对其它几个类有依赖，必须依赖其它几个类才能创建。

下面代码就是初始化得到DbBTree实例的一个过程。

       /// <summary>/// 数据库目录/// </summary>        private string directory;/// <summary>/// 数据库文件名/// </summary>        private string dbName;

private DbBTree btree;private Txn txn;private Db db;private Env env;/// <summary>/// 初始化/// </summary>        private void Init()        {            env = new Env(EnvCreateFlags.None);            Env.OpenFlags envFlags =              Env.OpenFlags.Create |              Env.OpenFlags.InitLock |              Env.OpenFlags.InitLog |              Env.OpenFlags.InitMPool |              Env.OpenFlags.InitTxn |              Env.OpenFlags.Recover;            env.Open(directory, envFlags, 0);            txn = env.TxnBegin(null, Txn.BeginFlags.None);            db = env.CreateDatabase(DbCreateFlags.None);            btree = (DbBTree)db.Open(txn, dbName, null, DbType.BTree, Db.OpenFlags.Create, 0);        }

另外Berkeley DB数据库的操作需要借助于序列化。

       /// <summary>/// 二进制序列化/// </summary>        private BinaryFormatter formatter;/// <summary>/// 键内存流/// </summary>        private MemoryStream keyStream;/// <summary>/// 内容内存流/// </summary>        private MemoryStream dataStream;

private void StreamInit()        {            formatter = new BinaryFormatter();            keyStream = new MemoryStream();            dataStream = new MemoryStream();        }

Berkeley DB是键值数据库，因此定义一个获取键接口：

    public interface IPut    {string Key { get; }    }

一、数据库的保存与更新

        public bool Set(IPut put)        {            Reset();

            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            dataStream.Position = 0;            formatter.Serialize(dataStream, put);            DbEntry data = DbEntry.InOut(dataStream.GetBuffer(), 0, (int)dataStream.Position);            WriteStatus status = btree.Put(txn, ref key, ref data);switch (status)            {case WriteStatus.Success:return true;case WriteStatus.NotFound:case WriteStatus.KeyExist:default:return false;            }        }

上述代码就可以保存键值。显示对键值进行序列化，然后再保存。保存完有三个状态，可以一一处理。

二、数据库的删除

删除最为简单

        public bool Remove(IPut put)        {            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            DeleteStatus status = btree.Delete(txn, ref key);switch (status)            {case DeleteStatus.NotFound:case DeleteStatus.Success:return true;case DeleteStatus.KeyEmpty:default:return false;            }        }

三、关于添加和删除

添加和删除并没有真正得进行添加和删除。必须执行Commit操作：

        private bool iscomit = false;public void Commit()        {            txn.Commit(Txn.CommitMode.None);            iscomit = true;        }

四、寻找键

用键查询值，和hash表一样使用。

        public bool Get(ref IPut put)        {            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            dataStream.SetLength(dataStream.Capacity);            DbEntry data = DbEntry.Out(dataStream.GetBuffer());

while (true)            {                ReadStatus status = btree.Get(txn, ref key, ref data, DbFile.ReadFlags.None);

switch (status)                {case ReadStatus.Success:                        dataStream.Position = 0;                        dataStream.SetLength(data.Size);                        put = (IPut)formatter.Deserialize(dataStream);return true;case ReadStatus.BufferSmall: //扩容                        if (key.Buffer.Length < key.Size)                        {                            keyStream.SetLength(key.Size);                            key = DbEntry.Out(keyStream.GetBuffer());                        }if (data.Buffer.Length < data.Size)                        {                            dataStream.SetLength(data.Size);                            data = DbEntry.Out(dataStream.GetBuffer());                        }continue;case ReadStatus.NotFound:case ReadStatus.KeyEmpty:default:return false;                }            }        }

五、遍历

public List<IPut> Find()        {            List<IPut> custList = new List<IPut>();using (DbBTreeCursor cursor = btree.OpenCursor(txn, DbFileCursor.CreateFlags.None))            {                IPut cust = null;while (GetNextRecord(cursor, ref cust))                    custList.Add(cust);            }return custList;        }

private bool GetNextRecord(DbBTreeCursor cursor, ref IPut cust)        {            ReadStatus status;            keyStream.SetLength(keyStream.Capacity);            dataStream.SetLength(dataStream.Capacity);            DbEntry key = DbEntry.Out(keyStream.GetBuffer());            DbEntry data = DbEntry.Out(dataStream.GetBuffer());do            {                status = cursor.Get(ref key, ref data, DbFileCursor.GetMode.Next, DbFileCursor.ReadFlags.None);switch (status)                {case ReadStatus.NotFound: return false;case ReadStatus.KeyEmpty: continue;  // skip deleted records                    case ReadStatus.BufferSmall:if (key.Buffer.Length < key.Size)                        {                            keyStream.SetLength(key.Size);                            key = DbEntry.Out(keyStream.GetBuffer());                        }if (data.Buffer.Length < data.Size)                        {                            dataStream.SetLength(data.Size);                            data = DbEntry.Out(dataStream.GetBuffer());                        }continue;case ReadStatus.Success:                        dataStream.Position = 0;                        dataStream.SetLength(data.Size);                        cust = (IPut)formatter.Deserialize(dataStream);return true;default:return false;                }            } while (true);        }

六、完整操作封装

代码：BerkeleyDB操作封装    public interface IPut    {string Key { get; }    }

public class BDBManager : IDisposable    {/// <summary>/// 数据库目录/// </summary>        private string directory;/// <summary>/// 数据库文件名/// </summary>        private string dbName;

private DbBTree btree;private Txn txn;private Db db;private Env env;

/// <summary>/// 二进制序列化/// </summary>        private BinaryFormatter formatter;/// <summary>/// 键内存流/// </summary>        private MemoryStream keyStream;/// <summary>/// 内容内存流/// </summary>        private MemoryStream dataStream;

public BDBManager(string directory, string dbName)        {this.directory = directory;this.dbName = dbName;

            Init();            StreamInit();        }

public bool Set(IPut put)        {            Reset();

            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            dataStream.Position = 0;            formatter.Serialize(dataStream, put);            DbEntry data = DbEntry.InOut(dataStream.GetBuffer(), 0, (int)dataStream.Position);            WriteStatus status = btree.Put(txn, ref key, ref data);switch (status)            {case WriteStatus.Success:return true;case WriteStatus.NotFound:case WriteStatus.KeyExist:default:return false;            }        }

private bool iscomit = false;public void Commit()        {            txn.Commit(Txn.CommitMode.None);            iscomit = true;        }

public List<IPut> Find()        {            List<IPut> custList = new List<IPut>();using (DbBTreeCursor cursor = btree.OpenCursor(txn, DbFileCursor.CreateFlags.None))            {                IPut cust = null;while (GetNextRecord(cursor, ref cust))                    custList.Add(cust);            }return custList;        }

public bool Get(ref IPut put)        {            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            dataStream.SetLength(dataStream.Capacity);            DbEntry data = DbEntry.Out(dataStream.GetBuffer());

while (true)            {                ReadStatus status = btree.Get(txn, ref key, ref data, DbFile.ReadFlags.None);

switch (status)                {case ReadStatus.Success:                        dataStream.Position = 0;                        dataStream.SetLength(data.Size);                        put = (IPut)formatter.Deserialize(dataStream);return true;case ReadStatus.BufferSmall: //扩容                        if (key.Buffer.Length < key.Size)                        {                            keyStream.SetLength(key.Size);                            key = DbEntry.Out(keyStream.GetBuffer());                        }if (data.Buffer.Length < data.Size)                        {                            dataStream.SetLength(data.Size);                            data = DbEntry.Out(dataStream.GetBuffer());                        }continue;case ReadStatus.NotFound:case ReadStatus.KeyEmpty:default:return false;                }            }        }

public bool Remove(IPut put)        {            Reset();

            keyStream.Position = 0;            formatter.Serialize(keyStream, put.Key);            DbEntry key = DbEntry.InOut(keyStream.GetBuffer(), 0, (int)keyStream.Position);            DeleteStatus status = btree.Delete(txn, ref key);switch (status)            {case DeleteStatus.NotFound:case DeleteStatus.Success:return true;case DeleteStatus.KeyEmpty:default:return false;            }        }

public void Dispose()        {if (!iscomit)                Commit();            db.Close();            db.Close();        }

private void Reset()        {            iscomit = false;        }

private void Init()        {            env = new Env(EnvCreateFlags.None);            Env.OpenFlags envFlags =              Env.OpenFlags.Create |              Env.OpenFlags.InitLock |              Env.OpenFlags.InitLog |              Env.OpenFlags.InitMPool |              Env.OpenFlags.InitTxn |              Env.OpenFlags.Recover;            env.Open(directory, envFlags, 0);            txn = env.TxnBegin(null, Txn.BeginFlags.None);            db = env.CreateDatabase(DbCreateFlags.None);            btree = (DbBTree)db.Open(txn, dbName, null, DbType.BTree, Db.OpenFlags.Create, 0);        }

private void StreamInit()        {            formatter = new BinaryFormatter();            keyStream = new MemoryStream();            dataStream = new MemoryStream();        }

private bool GetNextRecord(DbBTreeCursor cursor, ref IPut cust)        {            ReadStatus status;            keyStream.SetLength(keyStream.Capacity);            dataStream.SetLength(dataStream.Capacity);            DbEntry key = DbEntry.Out(keyStream.GetBuffer());            DbEntry data = DbEntry.Out(dataStream.GetBuffer());do            {                status = cursor.Get(ref key, ref data, DbFileCursor.GetMode.Next, DbFileCursor.ReadFlags.None);switch (status)                {case ReadStatus.NotFound: return false;case ReadStatus.KeyEmpty: continue;  // skip deleted records                    case ReadStatus.BufferSmall:if (key.Buffer.Length < key.Size)                        {                            keyStream.SetLength(key.Size);                            key = DbEntry.Out(keyStream.GetBuffer());                        }if (data.Buffer.Length < data.Size)                        {                            dataStream.SetLength(data.Size);                            data = DbEntry.Out(dataStream.GetBuffer());                        }continue;case ReadStatus.Success:                        dataStream.Position = 0;                        dataStream.SetLength(data.Size);                        cust = (IPut)formatter.Deserialize(dataStream);return true;default:return false;                }            } while (true);        }    }

调用方法：

首先要有一个写入的实体类，必须可以序列化,并且实现IPut接口：

        [Serializable()]class Item : IPut        {public string Name { get; set; }public string Text { get; set; }public int ID { get; set; }

public override string ToString()            {return string.Format("ID:{0} Key:{1}", ID, Name);            }

public string Key            {get { return Name; }            }        }

操作：

           using (BDBManager manager = new BDBManager("db", "db.dat"))            {bool success = manager.Set(new Item() { ID = 1000, Name = "Test",Text = "213" });                Console.WriteLine(string.Format("set is {0}", success));            }

using (BDBManager manager = new BDBManager("db", "db.dat"))            {                IPut put =  new Item() { Name = "Test" };bool success = manager.Get(ref put);                Console.WriteLine(string.Format("read is {0},item : {1}", success, put.ToString()));            }

using (BDBManager manager = new BDBManager("db", "db.dat"))            {                IPut put = new Item() { Name = "Test" };bool success = manager.Remove(put);                Console.WriteLine(string.Format("remove is {0},item : {1}", success, put.ToString()));            }

using (BDBManager manager = new BDBManager("db", "db.dat"))            {                List<IPut> list = manager.Find();foreach (var item in list)                {                    Console.WriteLine(item.ToString());                }            }            Console.WriteLine("end");            Console.ReadKey();

C#使用BerkeleyDB操作简介相关推荐

python中常见的流程结构-Python分支结构（switch）操作简介
Python当中并无switch语句,本文研究的主要是通过字典实现switch语句的功能,具体如下. switch语句用于编写多分支结构的程序,类似与if-.elif-.else语句. switch语 ...
python分支switch_Python分支结构（switch）操作简介
Python当中并无switch语句,本文研究的主要是通过字典实现switch语句的功能,具体如下. switch语句用于编写多分支结构的程序,类似与if-.elif-.else语句. switch语 ...
python中分支结构包括哪些_Python分支结构（switch）操作简介
Python当中并无switch语句,本文研究的主要是通过字典实现switch语句的功能,具体如下. switch语句用于编写多分支结构的程序,类似与if-.elif-.else语句. switch语 ...
oracle kepserve,Kepware.KEPServer\KEPServerEX_V5操作简介含opc quick client 连接测试
[实例简介] Kepware.KEPServer\KEPServerEX_V5操作简介含opc quick client 连接测试 3.開魯 KEPServereX的管理視窗請從「開始」->「 ...
作业一：数据软件系统的介绍及spss的操作简介
数据分析软件介绍数据分析是指用适当的统计分析方法对收集来的大量数据进行分析,提取有用信息和形成结论而对数据加以详细研究和概括总结的过程.这一过程也是质量管理体系的支持过程.在实用中, ...
JMeter入门操作简介
JMeter入门操作简介 1.JMter简介及作用 JMter是Apache公司用Java开发的一款测试工具接口测试.性能测试.数据库测试.java程序测试. 可以对服务器.网络或者对象模拟巨大的负 ...
UE官方教程笔记03-功能、术语、操作简介
对官方教程视频[官方培训]03-UE功能.术语.操作简介 | 徐良安 Epic的笔记这一部分基本都是走马观花的简单介绍功能世界创建建模 Mesh editingtool是一个全新的建模工具,具 ...
stm32 HAL库 Flash操作简介
stm32 HAL库 Flash操作简介目录第一stm32,flash介绍查看代码段,以判断代码长度 flash的基本操作规则 stm32 HAL库 Flash操作指南 stm32f1xx_ha ...
Python|并发编程|爬虫|单线程|多线程|异步I/O|360图片|Selenium及JavaScript|Scrapy框架|BOM 和 DOM 操作简介|语言基础50课：学习（12）
文章目录系列目录原项目地址第37课:并发编程在爬虫中的应用单线程版本多线程版本异步I/O版本总结第38课:抓取网页动态内容 Selenium 介绍使用Selenium 加载页面查找 ...

C#使用BerkeleyDB操作简介

C#使用BerkeleyDB操作简介相关推荐

最新文章

热门文章