(row:string, column:string, time:int64)→string

table {
  "1" : {"v1"},
  "2" : {"v2"},
  ...
}

student {
  "1" : {
    "info:name" : {"Tom"},
    "info:age" : {"20"},
    "grade:math" : {"100"}
  },
  "2" : {
    "info:name" : {"Jerry"},
    "info:age" : {"21"},
    "grade:math" : {"90"}
  },  
}

student {
  "1" : {
    "info" : {
      "name" : {"Tom"},
      "age" : {"20"}
    },
    "grade" : {
      "math" : "100"
    }
  }
}

student {
  "1" : {
    "info:name" : {
        1 : "Tom
    },
    "info:age" : {
        1 : "20"
    },
    "grade:math" : {
        7 : "90",
        10 : "100"
    }
  },
}

select * from tab 
where customer_name = 'Miller' and material = 'Refrigerator';

tar -zxvf hbase-1.3.1-bin.tar.gz -C ~/training/

HBASE_HOME=/root/training/hbase-1.3.1
export HBASE_HOME
PATH=$HBASE_HOME/bin:$HIVE_HOME/bin:$PATH
export PATH

source ~/.bash_profile

...
# Set environment variables here.

# This script sets variables multiple times over the course of starting an hbase process,
# so try to keep things idempotent unless you want to take an even deeper look
# into the startup scripts (bin/hbase, etc.)

# The java implementation to use.  Java 1.7+ required.
export JAVA_HOME=/root/training/jdk1.8.0_181
...

<configuration>
    <!--数据存储位置-->
    <property>
        <name>hbase.rootdir</name>
        <!--数据保存在本地-->
        <value>file:///root/training/hbase-1.3.1/data</value>
    </property>
</configuration>

start-hbase.sh

# 进入 hbase shell
[root@bigdata111 bin]# hbase shell
...
# 创建一个 student 表，表中有两个列族 info、grade
hbase(main):001:0> create 'student','info','grade'
...
# 查看库中有哪些表
hbase(main):002:0> list
...
# 往表中插入数据
hbase(main):003:0> put 'student','s01','info:name','Tom'
...
# 查看表中数据
hbase(main):004:0> scan 'student'
...
# 退出
hbase(main):005:0> exit

stop-hbase.sh

export HBASE_MANAGES_ZK=true

<configuration>
    <!--数据存储位置-->
    <property>
        <name>hbase.rootdir</name>
        <!--数据保存在HDFS上-->
        <value>hdfs://bigdata111:9000/hbase</value>
    </property>

    <!--表示是一个分布式的环境-->
    <property>
        <name>hbase.cluster.distributed</name>
        <value>true</value>
    </property>  

    <!--ZK的地址-->
    <property>
        <name>hbase.zookeeper.quorum</name>
        <value>bigdata111</value>
    </property>

    <!--Region的冗余-->
    <property>
        <name>dfs.replication</name>
        <value>1</value>
    </property>  
</configuration>

bigdata111

start-all.sh

start-hbase.sh

HBASE_HOME=/root/training/hbase-1.3.1
export HBASE_HOME

...
export JAVA_HOME=/root/training/jdk1.8.0_181
...
export HBASE_MANAGES_ZK=true
...

<configuration>
    <!--数据存储位置-->
    <property>
        <name>hbase.rootdir</name>
        <!--数据保存在HDFS上-->
        <value>hdfs://bigdata111:9000/hbase</value>
    </property>

    <!--表示是一个分布式的环境-->
    <property>
        <name>hbase.cluster.distributed</name>
        <value>true</value>
    </property>

    <!--ZK的地址-->
    <property>
        <name>hbase.zookeeper.quorum</name>
        <value>bigdata111</value>
    </property>

    <!--Region的冗余-->
    <property>
        <name>dfs.replication</name>
        <value>2</value>
    </property>
</configuration>

bigdata112
bigdata113

scp -r /root/training/hbase-1.3.1/ root@bigdata112:/root/training
scp -r /root/training/hbase-1.3.1/ root@bigdata113:/root/training

start-all.sh
start-hbase.sh

# bigdata111
[root@bigdata111 conf]# jps
2338 NameNode
2530 SecondaryNameNode
2692 ResourceManager
5013 HQuorumPeer
5290 Jps
5084 HMaster

# bigdata112
[root@bigdata112 conf]# jps
1461 NodeManager
1941 Jps
1350 DataNode

# bigdata113
[root@bigdata113 conf]# jps
1880 Jps
1307 DataNode
1419 NodeManager

2022-03-22 00:11:51,939 INFO  [bigdata111:16000.activeMasterManager] master.ServerManager: Waiting for region servers count to settle; currently checked in 0, slept for 428963 ms, expecting minimum of 1, maximum of 2147483647, timeout of 4500 ms, interval of 1500 ms.
2022-03-22 00:11:53,461 INFO  [bigdata111:16000.activeMasterManager] master.ServerManager: Waiting for region servers count to settle; currently checked in 0, slept for 430485 ms, expecting minimum of 1, maximum of 2147483647, timeout of 4500 ms, interval of 1500 ms.

# 同步方法1：所有机器直接同步到某个时间
date -s "yyyy-MM-dd HH:mm:ss"

# 同步方法2：跟时间服务器同步
ntpdate [timeserver]

# 进入 hbase zkcli
hbase zkcli
...
# 查看存在的表信息
[zk: bigdata111:2181(CONNECTED) 0] ls /hbase/table
[hbase:meta, hbase:namespace, student]
# 删除表信息
[zk: bigdata111:2181(CONNECTED) 1] rmr /hbase/table/student
# 再次查看是否删除成功
[zk: bigdata111:2181(CONNECTED) 2] ls /hbase/table
[hbase:meta, hbase:namespace]
# 退出
[zk: bigdata111:2181(CONNECTED) 3] quit

hbase-daemon.sh start master

hbase-daemon.sh start regionserver
hbase-daemon.sh start master

# 语法：create 'tablename','column-family1','column-family2',...,'column-familyN'
#
# 创建一个 student 表，表中有两个列族 info、grade
create 'student','info','grade'

# 语法：list
#
# 相当于SQL中 show tables 的效果
hbase(main):001:0> list

# 也可以使用缩写 desc '表名'
hbase(main):003:0> describe 'student'
Table student is ENABLED                                                                                                                                                                
student                                                                                                                                                                                 
COLUMN FAMILIES DESCRIPTION                                                                                                                                                             
{NAME => 'grade', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESSION => 'NONE', MI
N_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}                                                                                                
{NAME => 'info', BLOOMFILTER => 'ROW', VERSIONS => '1', IN_MEMORY => 'false', KEEP_DELETED_CELLS => 'FALSE', DATA_BLOCK_ENCODING => 'NONE', TTL => 'FOREVER', COMPRESSION => 'NONE', MIN
_VERSIONS => '0', BLOCKCACHE => 'true', BLOCKSIZE => '65536', REPLICATION_SCOPE => '0'}                                                                                                 
2 row(s) in 0.0860 seconds

# 语法：put 'tablename','rawkey','family1:column1','value1',...,'familyN:columnN','valueN'
#
# 往学生表插入一行数据，rawkey 为 s001，info列族的name属性为 Tom
hbase(main):005:0> put 'student','s001','info:name','Tom'
0 row(s) in 0.0550 seconds

# 语法：scan 'tablename'
#
# 查询学生表中所有数据，相当于：select * from student;
hbase(main):006:0> scan 'student'
ROW                                COLUMN+CELL                                                  s001                               column=info:name, timestamp=1648041822338, value=Tom          1 row(s) in 0.0290 seconds

# 语法：get 'tablename','rawkey'
#
# 查询 student 表中 rawkey 为 's001' 的数据
hbase(main):007:0> get 'student','s001'
COLUMN                                          CELL                                              info:name                                      timestamp=1648041822338, value=Tom               1 row(s) in 0.0140 seconds

hbase(main):032:0> get 'student','s001'
COLUMN                                         CELL                                             info:name                                      timestamp=1648043400032, value=Jerry             1 row(s) in 0.0040 seconds

hbase(main):033:0> get 'student','s001',{ COLUMN => 'info:name', TIMESTAMP => 1648043355198 }
COLUMN                                         CELL                                             info:name                                      timestamp=1648043355198, value=Tom               1 row(s) in 0.0060 seconds

hbase(main):034:0> get 'student','s001',{ COLUMN => 'info:name', TIMESTAMP => 1648043400032 }
COLUMN                                         CELL                                              info:name                                      timestamp=1648043400032, value=Jerry             1 row(s) in 0.0080 seconds

# 语法：delete 'tablename','rawkey','family:column'
#
# 删除数据不是以行为单位的，而是列
hbase(main):048:0> delete 'student','s001','grade'
0 row(s) in 0.0030 seconds

hbase(main):049:0> get 'student','s001'
COLUMN                                         CELL                                              grade:math                                     timestamp=1648044130996, value=100               grade:music                                    timestamp=1648044154962, value=95                 info:name                                      timestamp=1648043400032, value=Jerry             1 row(s) in 0.0030 seconds

hbase(main):050:0> delete 'student','s001','grade:math'
0 row(s) in 0.0030 seconds

hbase(main):051:0> get 'student','s001'
COLUMN                                         CELL                                              grade:music                                    timestamp=1648044154962, value=95                 info:name                                      timestamp=1648043400032, value=Jerry             1 row(s) in 0.0070 seconds

hbase(main):052:0> deleteall 'student','s001'
0 row(s) in 0.0080 seconds

hbase(main):053:0> get 'student','s001'
COLUMN                                          CELL                                                                                                                                    
0 row(s) in 0.0040 seconds

# 语法：drop 'tablename'
#
# 删除表前必须先禁用表，否则会报错：ERROR: Table student is enabled. Disable it first.
hbase(main):024:0> disable 'student'
0 row(s) in 2.2630 seconds

hbase(main):025:0> drop 'student'
0 row(s) in 1.2520 seconds

    // 如果程序执行卡住不动，可能是客户端无法解析 hbase 主机名，检查 host 即可
    // 创建表
    @Test
    public void testCreateTable() throws Exception {
        // 配置ZooKeeper的地址
        Configuration conf = new Configuration();
        conf.set("hbase.zookeeper.quorum", "192.168.190.111");
        
        // 客户端
        HBaseAdmin admin = new HBaseAdmin(conf);
        
        //创建表的描述符
        HTableDescriptor htd = new HTableDescriptor(TableName.valueOf("mystudent"));
        // 表中的列族
        htd.addFamily(new HColumnDescriptor("info"));
        htd.addFamily(new HColumnDescriptor("grade"));
        
        // 创建表
        admin.createTable(htd);
        
        admin.close();
    }

    // 插入数据:  单条数据
    @Test
    public void testPut() throws Exception {
        // 配置ZooKeeper的地址
        Configuration conf = new Configuration();
        conf.set("hbase.zookeeper.quorum", "192.168.190.111");
        
        // 得到表的客户端
        HTable table = new HTable(conf, "mystudent");
        
        // 插入的数据：构造Put(参数：行键)
        Put put = new Put(Bytes.toBytes("s001"));
        // put.addColumn(family,        列族的名字
        //               qualifier,     列的名字
        //               value)         值
        put.addColumn(Bytes.toBytes("info"), 
                      Bytes.toBytes("name"), 
                      Bytes.toBytes("Tom"));
        table.put(put);
        
        // 插入数据：多条数据
        // table.put(List<Put>);
        table.close();  
    }

    // 查询数据Get
    @Test
    public void testGet() throws Exception {
        // 配置ZooKeeper的地址
        Configuration conf = new Configuration();
        conf.set("hbase.zookeeper.quorum", "192.168.190.111");
        
        // 得到表的客户端
        HTable table = new HTable(conf, "mystudent");
        
        // 构造Get对象
        Get get = new Get(Bytes.toBytes("s001"));
        
        // 查询
        Result r = table.get(get);
        // 取出数据
        String name = Bytes.toString(r.getValue(Bytes.toBytes("info"), Bytes.toBytes("name")));
        System.out.println(name);
        
        table.close();
    }

    // 查询数据
    @Test
    public void testScan() throws Exception {
        // 配置ZooKeeper的地址
        Configuration conf = new Configuration();
        conf.set("hbase.zookeeper.quorum", "192.168.190.111");
        
        // 得到表的客户端
        HTable table = new HTable(conf, "mystudent");
        
        // 定义一个扫描器
        Scan scan = new Scan();
        // scan.setFilter(filter) 过滤器
        
        // 通过扫描器查询数据
        // 在该集合中保存的是Result
        ResultScanner rs = table.getScanner(scan);
        for(Result r:rs) {
            //取出数据
            String name = Bytes.toString(r.getValue(Bytes.toBytes("info"), Bytes.toBytes("name")));
            System.out.println(name);           
        }
        
        table.close();
    }

    // 删除表
    @Test
    public void testDrop() throws Exception {
        // 配置ZooKeeper的地址
        Configuration conf = new Configuration();
        conf.set("hbase.zookeeper.quorum", "192.168.190.111");
        
        // 客户端
        HBaseAdmin admin = new HBaseAdmin(conf);
        
        // 删除表
        admin.disableTable("mystudent");
        admin.deleteTable("mystudent");
        
        admin.close();
    }

public class WordCountDataInit {
@Test
public void createTable() throws IOException {
//指定的配置信息: ZooKeeper
Configuration conf = new Configuration();
conf.set("hbase.zookeeper.quorum", "192.168.190.111");
//创建一个HBase客户端: HBaseAdmin
HBaseAdmin admin = new HBaseAdmin(conf);
//创建一个表的描述符: 表名
HTableDescriptor hd = new HTableDescriptor(TableName.valueOf("word"));
//创建列族描述符
HColumnDescriptor hcd1 = new HColumnDescriptor("content");
//加入列族
hd.addFamily(hcd1);
//创建表
admin.createTable(hd);
//关闭客户端
admin.close();
}
@Test
public void insertData() throws IOException {
//指定的配置信息: ZooKeeper
Configuration conf = new Configuration();
conf.set("hbase.zookeeper.quorum", "192.168.190.111");
//客户端
HTable table = new HTable(conf, "word");
Put put1 = new Put(Bytes.toBytes("1"));
put1.add(Bytes.toBytes("content"), Bytes.toBytes("info"), Bytes.toBytes("I love Beijing"));
Put put2 = new Put(Bytes.toBytes("2"));
put2.add(Bytes.toBytes("content"), Bytes.toBytes("info"), Bytes.toBytes("I love China"));
Put put3 = new Put(Bytes.toBytes("3"));
put3.add(Bytes.toBytes("content"), Bytes.toBytes("info"), Bytes.toBytes("Beijing is the capital of China"));
List<Put> list = new ArrayList<Put>();
list.add(put1);
list.add(put2);
list.add(put3);
//插入数据
table.put(list);
table.close();  
}
}

// 现在的输入就是HBase表中的一条记录
//                                              k2单词    v2记一次数
public class WordCountMapper extends TableMapper<Text, IntWritable> {
@Override
protected void map(ImmutableBytesWritable key1, Result value1, Context context)
throws IOException, InterruptedException {
/*
* key1：代表的是行键rowkey
* value1：该行记录
* 数据 : I love Beijing
*/
String data = Bytes.toString(value1.getValue(Bytes.toBytes("content"), Bytes.toBytes("info")));
// 分词
String[] words = data.split(" ");
for (String word : words) {
context.write(new Text(word), new IntWritable(1));
}
}
}

//                                                  k3       v3           k4是输出记录的行键
public class WordCountReducer extends TableReducer<Text, IntWritable, ImmutableBytesWritable> {
@Override
protected void reduce(Text key3, Iterable<IntWritable> value3, Context context)
throws IOException, InterruptedException {
// 计算总数
int total = 0;
for (IntWritable v : value3) {
total += v.get();
}
// 输出的是表中的一条记录
// 构造一个Put对象
Put put = new Put(Bytes.toBytes(key3.toString()));
put.addColumn(Bytes.toBytes("content"), Bytes.toBytes("result"), Bytes.toBytes(String.valueOf(total)));
// 输出                单词作为rowkey
context.write(new ImmutableBytesWritable(Bytes.toBytes(key3.toString())), put);
}
}

public class WordCountMain {
public static void main(String[] args) throws Exception {
// 指定的配置信息: ZooKeeper
Configuration conf = new Configuration();
conf.set("hbase.zookeeper.quorum", "192.168.190.111");
// 创建任务，指定任务的入口
Job job = Job.getInstance(conf);
job.setJarByClass(WordCountMain.class);
// 定义一个扫描器，指定读取的列
Scan scan = new Scan();
scan.addColumn(Bytes.toBytes("content"), Bytes.toBytes("info"));
// 使用工具类指定任务的Map
TableMapReduceUtil.initTableMapperJob(Bytes.toBytes("word"), // 指定读取HBase数据所在的表
scan,                // 扫描器
WordCountMapper.class,// 指定处理的Mapper Class
Text.class,           // k2
IntWritable.class,    // v2
job);
// 使用工具类指定任务的reducer
TableMapReduceUtil.initTableReducerJob("stat", 
WordCountReducer.class, 
job);
// 执行任务
job.waitForCompletion(true);
}
}

hadoop jar hbase-mr.jar