https://blog.youkuaiyun.com/jackfor001zh-cnhttps://blog.youkuaiyun.com/https://blog.youkuaiyun.com/jackfor001/article/details/8063698https://blog.youkuaiyun.com/jackfor001/article/details/8063698jackfor001Fri, 12 Oct 2012 10:54:46 +0800https://blog.youkuaiyun.com/jackfor001/article/details/7225037https://blog.youkuaiyun.com/jackfor001/article/details/7225037jackfor001Wed, 01 Feb 2012 11:36:04 +0800https://blog.youkuaiyun.com/jackfor001/article/details/7174237https://blog.youkuaiyun.com/jackfor001/article/details/7174237jackfor001Wed, 04 Jan 2012 10:44:35 +0800https://blog.youkuaiyun.com/jackfor001/article/details/7031943https://blog.youkuaiyun.com/jackfor001/article/details/7031943jackfor001Thu, 01 Dec 2011 18:13:24 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6757989https://blog.youkuaiyun.com/jackfor001/article/details/6757989jackfor001Wed, 07 Sep 2011 20:20:11 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6752047https://blog.youkuaiyun.com/jackfor001/article/details/6752047jackfor001Tue, 06 Sep 2011 08:44:33 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6561265https://blog.youkuaiyun.com/jackfor001/article/details/6561265jackfor001Wed, 22 Jun 2011 15:36:00 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6394683https://blog.youkuaiyun.com/jackfor001/article/details/6394683jackfor001Wed, 04 May 2011 17:35:00 +0800社交型网络（SNS, Social Networking Services）以六度分割理论[1]为基础，构建好友圈并通过“好友的好友”来无限扩张人脉。社交网络在近几年获得了空前的发展，像国内的校内网，开心网已经聚集了几千万的用户。SNS 特别之处在于维护朋友关系，增进友谊，并拓展交际圈，以用户之间的黏性来保持用户和SNS 的黏性。用户虽然也可以设置过滤条件搜索得到符合特定条件的用户群组，但是由于庞大的用户基数，在通常情况下搜索得出的用户群组其中大部分对特定用户来说都是冗余的，而不是将与特定用户熟]]>https://blog.youkuaiyun.com/jackfor001/article/details/6193099https://blog.youkuaiyun.com/jackfor001/article/details/6193099jackfor001Fri, 18 Feb 2011 13:02:00 +0800
当map task开始运算，并产生中间数据时，其产生的中间结果并非直接就简单的写入磁盘。这中间的过程比较复杂，并且利用到了内存buffer来进行已经产生的部分结果的缓存，并在内存buffer中进行一些预排序来优化整个map的性能。如上图所示，每一个map都会对应存在一个内存buffer（MapOutputBuffer，即上图的buffer in memory），map会将已经产生的部分结果先写入到该buffer中]]>https://blog.youkuaiyun.com/jackfor001/article/details/6132019https://blog.youkuaiyun.com/jackfor001/article/details/6132019jackfor001Wed, 12 Jan 2011 17:00:00 +0800最近关注Hadoop，因此也顺便关注了一下Hadoop相关的项目。HBASE就是基于Hadoop的一个开源项目，也是对Google的BigTable的一种实现。
       BigTable是什么？Google的Paper对其作了充分的说明。字面上看就是一张大表，其实和我们想象的传统数据库的表还是有些差别的。松散数据可以说是介于Map Entry（key & value）和DB Row之间的一种数据。在我使用Memcache的时候，有时候的需求是需要存储的不仅仅是简单的一个key对]]>https://blog.youkuaiyun.com/jackfor001/article/details/6127900https://blog.youkuaiyun.com/jackfor001/article/details/6127900jackfor001Mon, 10 Jan 2011 21:57:00 +0800简单加载方式：
 
 
Merchant temp = new Merchant();//得到数据集
 
ReportViewer1.LocalReport.DataSources.Add(new Microsoft.Reporting.WebForms.ReportDataSource("Class1", temp.GreatList("")));//Class1是数据对象，类似DataSet
 
ReportViewer1]]>https://blog.youkuaiyun.com/jackfor001/article/details/6105477https://blog.youkuaiyun.com/jackfor001/article/details/6105477jackfor001Wed, 29 Dec 2010 17:15:00 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6092096https://blog.youkuaiyun.com/jackfor001/article/details/6092096jackfor001Wed, 22 Dec 2010 16:31:00 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6091769https://blog.youkuaiyun.com/jackfor001/article/details/6091769jackfor001Wed, 22 Dec 2010 14:53:00 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6088748https://blog.youkuaiyun.com/jackfor001/article/details/6088748jackfor001Tue, 21 Dec 2010 11:15:00 +0800https://blog.youkuaiyun.com/jackfor001/article/details/6082869https://blog.youkuaiyun.com/jackfor001/article/details/6082869jackfor001Fri, 17 Dec 2010 18:08:00 +0800HBase的许多代码是重用HADOOP的代码。它与我们常用的数据库的最大区别就是列存储以及无数据类型，所有数据都以String类型进行存储。而且，假如hbase table里面有5个字段，但是实际上只有4个字段有值，另一个为NULL，那么这个字段是不占存储空间的，相比较而言，这点比较好罢。
     HBase是一个开源的、分布式的、面向列的存储系统。源于论文《Bigtable：一个结构化数据的分布式存储系统》。就像Bigtable利用了Googl文件系统（File System）所提]]>https://blog.youkuaiyun.com/jackfor001/article/details/6082799https://blog.youkuaiyun.com/jackfor001/article/details/6082799jackfor001Fri, 17 Dec 2010 17:36:00 +0800一、安装准备
1、下载HBASE 0.20.5版本：http://www.apache.org/dist/hbase/hbase-0.20.5/
2、JDK版本：jdk-6u20-linux-i586.bin
3、操作系统：Linux s132 2.6.9-78.8AXS2smp #1 SMP Tue Dec 16 02:42:55 EST 2008 x86_64 x86_64 x86_64 GNU/Linux
4、默认前提是安装完hadoop 0.20]]>https://blog.youkuaiyun.com/jackfor001/article/details/6075678https://blog.youkuaiyun.com/jackfor001/article/details/6075678jackfor001Tue, 14 Dec 2010 16:10:00 +0800创建分表
 
CREATE TABLE tmp_table
(
passport string,
cid int
)PARTITIONED BY(vt bigint)--这个地方就是起个名字吗？
 ROW FORMAT DELIMITED
   FIELDS TERMINATED BY '/,'
STORED AS SEQUENCEFILE;
 
 
添加分区
 ]]>https://blog.youkuaiyun.com/jackfor001/article/details/6075556https://blog.youkuaiyun.com/jackfor001/article/details/6075556jackfor001Tue, 14 Dec 2010 15:33:00 +0800DDL Operations
创建表
hive> CREATE TABLE pokes (foo INT, bar STRING);
创建表并创建索引字段ds
hive> CREATE TABLE invites (foo INT, bar STRING) PARTITIONED BY (ds STRING);
显示所有表
hive> SHOW TABLES;
按正条件（正则表达式）显示表，
hive> SHOW]]>https://blog.youkuaiyun.com/jackfor001/article/details/6075336https://blog.youkuaiyun.com/jackfor001/article/details/6075336jackfor001Tue, 14 Dec 2010 14:30:00 +0800