在基于HBase数据库的开发中,对应Java语言来说,可以直接使用HBase的原生API来操作HBase表数据,当然你要是不嫌麻烦可以使用Thrift客户端Java API,这里有我曾经使用过的 HBase Thrift客户端Java API实践,可以参考。对于具有其他编程语言背景的开发人员,为
在基于HBase数据库的开发中,对应Java语言来说,可以直接使用HBase的原生API来操作HBase表数据,当然你要是不嫌麻烦可以使用Thrift客户端Java API,这里有我曾经使用过的 HBase Thrift客户端Java API实践,可以参考。对于具有其他编程语言背景的开发人员,为了获取HBase带来的好处,那么就可以选择使用HBase Thrift客户端对应编程语言的API,来实现与HBase的交互。
这里,我们使用C#客户端来操作HBase。HBase的Thrift接口的定义,可以通过链接http://svn.apache.org/viewvc/hbase/trunk/hbase-server/src/main/resources/org/apache/hadoop/hbase/thrift/Hbase.thrift?view=markup看到,我们需要安装Thrift编译器,才能生成HBase跨语言的API,这里,我使用的版本是0.9.0。需要注意的是,一定要保证,安装了某个版本Thrift的Thrift编译器,在导入对应语言库的时候,版本一定要统一,否则就会出现各种各样的问题,因为不同Thrift版本,对应编程语言的库API可能有变化。
首先,下载上面链接的内容,保存为Hbase.thrift。
然后,执行如下命令,生成C#编程语言的HBase Thrift客户端API:
[hadoop@master hbase]$ thrift --gen csharp Hbase.thrift [hadoop@master hbase]$ ls gen-csharp
这里,我们基于C#语言,使用HBase 的Thrift 客户端API访问HBase表。事实上,如果使用Java来实现对HBase表的操作,最好是使用HBase的原生API,无论从性能还是便利性方面,都会提供更好的体验。使用Thrift API访问,实际也是在HBase API之上进行了一层封装,可能初次使用Thrift API感觉很别扭,有时候还要参考Thrift服务端的实现代码。
准备工作如下:
- 下载Thrift软件包,解压缩后,拷贝thrift-0.9.0/lib/java/src下面的代码到工作区(开发工具中)
- 将上面生成的gen-csharp目录中代码拷贝到工作区
- 保证HBase集群正常运行,接着启动HBase的Thrift服务,执行如下命令:
bin/hbase thrift -b master -p 9090 start
上面,HBase的Thrift服务端口为9090,下面通过Thrift API访问的时候,需要用到,而不是HBase的服务端口(默认60000)。
接着,实现一个简单的例子,访问Hbase表。
首先,我们通过HBase Shell创建一个表:
create 'test_info', 'info'
表名为test_info,列簇名称为info。
然后,我们开始基于上面生成的Thrift代码来实现对HBase表的操作。
这里,我们实际上是对HBase Thrift客户端Java API实践中的Java代码进行了翻译,改写成C#语言的相关操作。我们在客户端,进行了一层抽象,更加便于传递各种参数,抽象类为AbstractHBaseThriftService,对应的命名空间为HbaseThrift.HBase.Thrift,该类实现代码如下所示:
using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; using Thrift.Transport; using Thrift.Protocol; namespace HbaseThrift.HBase.Thrift { public abstract class AbstractHBaseThriftService { protected static readonly string CHARSET = "UTF-8"; private string host = "localhost"; private int port = 9090; private readonly TTransport transport; protected readonly Hbase.Client client; public AbstractHBaseThriftService() : this("localhost", 9090) { } public AbstractHBaseThriftService(string host, int port) { this.host = host; this.port = port; transport = new TSocket(host, port); TProtocol protocol = new TBinaryProtocol(transport, true, true); client = new Hbase.Client(protocol); } public void Open() { if (transport != null) { transport.Open(); } } public void Close() { if (transport != null) { transport.Close(); } } public abstract List GetTables(); public abstract void Update(string table, string rowKey, bool writeToWal, string fieldName, string fieldValue, Dictionary attributes); public abstract void Update(string table, string rowKey, bool writeToWal, Dictionary fieldNameValues, Dictionary attributes); public abstract void DeleteCell(string table, string rowKey, bool writeToWal, string column, Dictionary attributes); public abstract void DeleteCells(string table, string rowKey, bool writeToWal, List columns, Dictionary attributes); public abstract void DeleteRow(string table, string rowKey, Dictionary attributes); public abstract int ScannerOpen(string table, string startRow, List columns, Dictionary attributes); public abstract int ScannerOpen(string table, string startRow, string stopRow, List columns, Dictionary attributes); public abstract int ScannerOpenWithPrefix(string table, string startAndPrefix, List columns, Dictionary attributes); public abstract int ScannerOpenTs(string table, string startRow, List columns, long timestamp, Dictionary attributes); public abstract int ScannerOpenTs(string table, string startRow, string stopRow, List columns, long timestamp, Dictionary attributes); public abstract List ScannerGetList(int id, int nbRows); public abstract List ScannerGet(int id); public abstract List GetRow(string table, string row, Dictionary attributes); public abstract List GetRows(string table, List rows, Dictionary attributes); public abstract List GetRowsWithColumns(string table, List rows, List columns, Dictionary attributes); public abstract void ScannerClose(int id); /** * Iterate result rows(just for test purpose) * @param result */ public abstract void IterateResults(TRowResult result); } }
这里,简单叙述一下,我们提供的客户端API的基本功能:
- 建立到Thrift服务的连接:Open()
- 获取到HBase中的所有表名:GetTables()
- 更新HBase表记录:Update()
- 删除HBase表中一行的记录的数据(cell):DeleteCell()和DeleCells()
- 删除HBase表中一行记录:deleteRow()
- 打开一个Scanner,返回id:ScannerOpen()、ScannerOpenWithPrefix()和ScannerOpenTs();然后用返回的id迭代记录:ScannerGetList()和ScannerGet()
- 获取一行记录结果:GetRow()、GetRows()和GetRowsWithColumns()
- 关闭一个Scanner:ScannerClose()
- 迭代结果,用于调试:IterateResults()
比如,我们想要实现分页的逻辑,可能和传统的关系型数据库操作有些不同。基于HBase表的实现是,首先打开一个Scanner实例(例如调用ScannerOpen()),返回一个id,然后再使用该id,调用ScannerGetList()方法(可以指定每次返回几条记录的变量nbRows的值),返回一个记录列表,反复调用该ScannerGetList()方法,直到此次没有结果返回为止。后面会通过测试用例来实际体会。
现在,我们基于上抽象出来的客户端操作接口,给出一个基本的实现,代码如下所示:
using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; namespace HbaseThrift.HBase.Thrift { class HBaseThriftClient : AbstractHBaseThriftService { public HBaseThriftClient() : this("localhost", 9090) { } public HBaseThriftClient(string host, int port) : base(host, port) { } public override List GetTables() { List tables = client.getTableNames(); List list = new List(); foreach(byte[] table in tables) { list.Add(Decode(table)); } return list; } public override void Update(string table, string rowKey, bool writeToWal, string fieldName, string fieldValue, Dictionary attributes) { byte[] tableName = Encode(table); byte[] row = Encode(rowKey); Dictionary encodedAttributes = EncodeAttributes(attributes); List mutations = new List(); Mutation mutation = new Mutation(); mutation.IsDelete = false; mutation.WriteToWAL = writeToWal; mutation.Column = Encode(fieldName); mutation.Value = Encode(fieldValue); mutations.Add(mutation); client.mutateRow(tableName, row, mutations, encodedAttributes); } public override void Update(string table, string rowKey, bool writeToWal, Dictionary fieldNameValues, Dictionary attributes) { byte[] tableName = Encode(table); byte[] row = Encode(rowKey); Dictionary encodedAttributes = EncodeAttributes(attributes); List mutations = new List(); foreach (KeyValuePair pair in fieldNameValues) { Mutation mutation = new Mutation(); mutation.IsDelete = false; mutation.WriteToWAL = writeToWal; mutation.Column = Encode(pair.Key); mutation.Value = Encode(pair.Value); mutations.Add(mutation); } client.mutateRow(tableName, row, mutations, encodedAttributes); } public override void DeleteCell(string table, string rowKey, bool writeToWal, string column, Dictionary attributes) { byte[] tableName = Encode(table); byte[] row = Encode(rowKey); Dictionary encodedAttributes = EncodeAttributes(attributes); List mutations = new List(); Mutation mutation = new Mutation(); mutation.IsDelete = true; mutation.WriteToWAL = writeToWal; mutation.Column = Encode(column); mutations.Add(mutation); client.mutateRow(tableName, row, mutations, encodedAttributes); } public override void DeleteCells(string table, string rowKey, bool writeToWal, List columns, Dictionary attributes) { byte[] tableName = Encode(table); byte[] row = Encode(rowKey); Dictionary encodedAttributes = EncodeAttributes(attributes); List mutations = new List(); foreach (string column in columns) { Mutation mutation = new Mutation(); mutation.IsDelete = true; mutation.WriteToWAL = writeToWal; mutation.Column = Encode(column); mutations.Add(mutation); } client.mutateRow(tableName, row, mutations, encodedAttributes); } public override void DeleteRow(string table, string rowKey, Dictionary attributes) { byte[] tableName = Encode(table); byte[] row = Encode(rowKey); Dictionary encodedAttributes = EncodeAttributes(attributes); client.deleteAllRow(tableName, row, encodedAttributes); } public override int ScannerOpen(string table, string startRow, List columns, Dictionary attributes) { byte[] tableName = Encode(table); byte[] start = Encode(startRow); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.scannerOpen(tableName, start, encodedColumns, encodedAttributes); } public override int ScannerOpen(string table, string startRow, string stopRow, List columns, Dictionary attributes) { byte[] tableName = Encode(table); byte[] start = Encode(startRow); byte[] stop = Encode(stopRow); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.scannerOpenWithStop(tableName, start, stop, encodedColumns, encodedAttributes); } public override int ScannerOpenWithPrefix(string table, string startAndPrefix, List columns, Dictionary attributes) { byte[] tableName = Encode(table); byte[] prefix = Encode(startAndPrefix); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.scannerOpenWithPrefix(tableName, prefix, encodedColumns, encodedAttributes); } public override int ScannerOpenTs(string table, string startRow, List columns, long timestamp, Dictionary attributes) { byte[] tableName = Encode(table); byte[] start = Encode(startRow); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.scannerOpenTs(tableName, start, encodedColumns, timestamp, encodedAttributes); } public override int ScannerOpenTs(string table, string startRow, string stopRow, List columns, long timestamp, Dictionary attributes) { byte[] tableName = Encode(table); byte[] start = Encode(startRow); byte[] stop = Encode(stopRow); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.scannerOpenWithStopTs(tableName, start, stop, encodedColumns, timestamp, encodedAttributes); } public override List ScannerGetList(int id, int nbRows) { return client.scannerGetList(id, nbRows); } public override List ScannerGet(int id) { return client.scannerGet(id); } public override List GetRow(string table, string row, Dictionary attributes) { byte[] tableName = Encode(table); byte[] startRow = Encode(row); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.getRow(tableName, startRow, encodedAttributes); } public override List GetRows(string table, List rows, Dictionary attributes) { byte[] tableName = Encode(table); List encodedRows = EncodeStringList(rows); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.getRows(tableName, encodedRows, encodedAttributes); } public override List GetRowsWithColumns(string table, List rows, List columns, Dictionary attributes) { byte[] tableName = Encode(table); List encodedRows = EncodeStringList(rows); List encodedColumns = EncodeStringList(columns); Dictionary encodedAttributes = EncodeAttributes(attributes); return client.getRowsWithColumns(tableName, encodedRows, encodedColumns, encodedAttributes); } public override void ScannerClose(int id) { client.scannerClose(id); } public override void IterateResults(TRowResult result) { foreach (KeyValuePair pair in result.Columns) { Console.WriteLine("\tCol=" + Decode(pair.Key) + ", Value=" + Decode(pair.Value.Value)); } } private String Decode(byte[] bs) { return UTF8Encoding.Default.GetString(bs); } private byte[] Encode(String str) { return UTF8Encoding.Default.GetBytes(str); } private Dictionary EncodeAttributes(Dictionary attributes) { Dictionary encodedAttributes = new Dictionary(); foreach (KeyValuePair pair in attributes) { encodedAttributes.Add(Encode(pair.Key), Encode(pair.Value)); } return encodedAttributes; } private List EncodeStringList(List strings) { List list = new List(); if (strings != null) { foreach (String str in strings) { list.Add(Encode(str)); } } return list; } } }
上面代码,给出了基本的实现,接着我们给出测试用例,调用我们实现的客户端操作,与HBase表进行交互。实现的测试用例类如下所示:
using System; using System.Collections.Generic; using System.Linq; using System.Text; using System.Threading.Tasks; namespace HbaseThrift.HBase.Thrift { class Test { private readonly AbstractHBaseThriftService client; public Test(String host, int port) { client = new HBaseThriftClient(host, port); } public Test() : this("master", 9090) { } static String RandomlyBirthday() { Random r = new Random(); int year = 1900 + r.Next(100); int month = 1 + r.Next(12); int date = 1 + r.Next(30); return year + "-" + month.ToString().PadLeft(2, '0') + "-" + date.ToString().PadLeft(2, '0'); } static String RandomlyGender() { Random r = new Random(); int flag = r.Next(2); return flag == 0 ? "M" : "F"; } static String RandomlyUserType() { Random r = new Random(); int flag = 1 + r.Next(10); return flag.ToString(); } public void Close() { client.Close(); } public void CaseForUpdate() { bool writeToWal = false; Dictionary attributes = new Dictionary(0); string table = SetTable(); // put kv pairs for (int i = 0; i <p>上面的测试可以实现操作Hbase表数据。另外,在生成的Thrift客户端代码中,Iface中给出了全部的服务接口,可以根据需要来选择,客户端Client实现了与Thrift交互的一些逻辑的处理,通过该类对象可以代理HBase提供的Thrift服务。</p> <p><strong>参考链接</strong></p>
- http://wiki.apache.org/hadoop/Hbase/ThriftApi
- http://svn.apache.org/viewvc/hbase/trunk/hbase-server/src/main/resources/org/apache/hadoop/hbase/thrift/Hbase.thrift?view=markup
- http://www.cnblogs.com/panfeng412/archive/2012/11/11/hbase-thrift-api-common-issues-summary.html
- https://github.com/simplegeo/hadoop-hbase/blob/master/src/examples/thrift/DemoClient.java
- http://thrift.apache.org/tutorial/java/
原文地址:基于C#+Thrift操作HBase实践, 感谢原作者分享。

ACID屬性包括原子性、一致性、隔離性和持久性,是數據庫設計的基石。 1.原子性確保事務要么完全成功,要么完全失敗。 2.一致性保證數據庫在事務前後保持一致狀態。 3.隔離性確保事務之間互不干擾。 4.持久性確保事務提交後數據永久保存。

MySQL既是數據庫管理系統(DBMS),也與編程語言緊密相關。 1)作為DBMS,MySQL用於存儲、組織和檢索數據,優化索引可提高查詢性能。 2)通過SQL與編程語言結合,嵌入在如Python中,使用ORM工具如SQLAlchemy可簡化操作。 3)性能優化包括索引、查詢、緩存、分庫分錶和事務管理。

MySQL使用SQL命令管理數據。 1.基本命令包括SELECT、INSERT、UPDATE和DELETE。 2.高級用法涉及JOIN、子查詢和聚合函數。 3.常見錯誤有語法、邏輯和性能問題。 4.優化技巧包括使用索引、避免SELECT*和使用LIMIT。

MySQL是一種高效的關係型數據庫管理系統,適用於存儲和管理數據。其優勢包括高性能查詢、靈活的事務處理和豐富的數據類型。實際應用中,MySQL常用於電商平台、社交網絡和內容管理系統,但需注意性能優化、數據安全和擴展性。

SQL和MySQL的關係是標準語言與具體實現的關係。 1.SQL是用於管理和操作關係數據庫的標準語言,允許進行數據的增、刪、改、查。 2.MySQL是一個具體的數據庫管理系統,使用SQL作為其操作語言,並提供高效的數據存儲和管理。

InnoDB使用redologs和undologs確保數據一致性和可靠性。 1.redologs記錄數據頁修改,確保崩潰恢復和事務持久性。 2.undologs記錄數據原始值,支持事務回滾和MVCC。

EXPLAIN命令的關鍵指標包括type、key、rows和Extra。 1)type反映查詢的訪問類型,值越高效率越高,如const優於ALL。 2)key顯示使用的索引,NULL表示無索引。 3)rows預估掃描行數,影響查詢性能。 4)Extra提供額外信息,如Usingfilesort提示需要優化。

Usingtemporary在MySQL查詢中表示需要創建臨時表,常見於使用DISTINCT、GROUPBY或非索引列的ORDERBY。可以通過優化索引和重寫查詢避免其出現,提升查詢性能。具體來說,Usingtemporary出現在EXPLAIN輸出中時,意味著MySQL需要創建臨時表來處理查詢。這通常發生在以下情況:1)使用DISTINCT或GROUPBY時進行去重或分組;2)ORDERBY包含非索引列時進行排序;3)使用複雜的子查詢或聯接操作。優化方法包括:1)為ORDERBY和GROUPB


熱AI工具

Undresser.AI Undress
人工智慧驅動的應用程序,用於創建逼真的裸體照片

AI Clothes Remover
用於從照片中去除衣服的線上人工智慧工具。

Undress AI Tool
免費脫衣圖片

Clothoff.io
AI脫衣器

AI Hentai Generator
免費產生 AI 無盡。

熱門文章

熱工具

VSCode Windows 64位元 下載
微軟推出的免費、功能強大的一款IDE編輯器

DVWA
Damn Vulnerable Web App (DVWA) 是一個PHP/MySQL的Web應用程序,非常容易受到攻擊。它的主要目標是成為安全專業人員在合法環境中測試自己的技能和工具的輔助工具,幫助Web開發人員更好地理解保護網路應用程式的過程,並幫助教師/學生在課堂環境中教授/學習Web應用程式安全性。 DVWA的目標是透過簡單直接的介面練習一些最常見的Web漏洞,難度各不相同。請注意,該軟體中

SublimeText3 Linux新版
SublimeText3 Linux最新版

Dreamweaver CS6
視覺化網頁開發工具

MantisBT
Mantis是一個易於部署的基於Web的缺陷追蹤工具,用於幫助產品缺陷追蹤。它需要PHP、MySQL和一個Web伺服器。請查看我們的演示和託管服務。