Kafka To HBase To Hive

1.在HBase中创建表

2.写入API

2.1普通模式写入hbase（逐条写入）

2.2普通模式写入hbase（buffer写入）

2.3设计模式写入hbase（buffer写入）

3.HBase表映射至Hive中

1.在HBase中创建表

hbase(main):003:0> create_namespace 'events_db'

hbase(main):004:0> create 'events_db:users','profile','region','registration'

hbase(main):005:0> create 'events_db:user_friend','uf'

hbase(main):006:0> create 'events_db:events','schedule','location','creator','remark'

hbase(main):007:0> create 'events_db:event_attendee','euat'

hbase(main):008:0> create 'events_db:train','eu'

hbase(main):011:0> list_namespace_tables 'events_db'

TABLE

event_attendee

events

train

user_friend

users

5 row(s)

2.写入API

2.1普通模式写入hbase（逐条写入）

java 复制代码

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.Connection;
import org.apache.hadoop.hbase.client.ConnectionFactory;
import org.apache.hadoop.hbase.client.Put;
import org.apache.hadoop.hbase.client.Table;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.kafka.clients.consumer.ConsumerConfig;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.common.serialization.StringDeserializer;

import java.io.IOException;
import java.time.Duration;

import java.util.ArrayList;
import java.util.Collections;
import java.util.Properties;

/**
 * 将Kafka中的topic为userfriends中的数据消费到hbase中
 * hbase中的表为events_db:user_friend
 */
public class UserFriendToHB {
    static int num = 0; //计数器
    public static void main(String[] args) {
        Properties properties = new Properties();
        properties.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, "kb129:9092");
        properties.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
        properties.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG,StringDeserializer.class);
        properties.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG,"earliest");
        properties.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG,"false");
        properties.put(ConsumerConfig.GROUP_ID_CONFIG, "user_friend_group1");

        KafkaConsumer<String, String> consumer = new KafkaConsumer<>(properties);
        consumer.subscribe(Collections.singleton("userfriends"));

        //配置hbase信息，连接hbase数据库
        Configuration conf = HBaseConfiguration.create();
        conf.set(HConstants.HBASE_DIR, "hdfs://kb129:9000/hbase");
        conf.set(HConstants.ZOOKEEPER_QUORUM, "kb129");
        conf.set(HConstants.CLIENT_PORT_STR, "2181");

        Connection connection = null;
        try {
            connection = ConnectionFactory.createConnection(conf);
            Table ufTable = connection.getTable(TableName.valueOf("events_db:user_friend"));
            ArrayList<Put> datas = new ArrayList<>();

            while (true){
                ConsumerRecords<String, String> poll = consumer.poll(Duration.ofMillis(100));
                //每次for循环前清空datas
                datas.clear();
                for (ConsumerRecord<String, String> record : poll) {
                    //System.out.println(record.value());
                    String[] split = record.value().split(",");
                    int i = (split[0] + split[1]).hashCode();
                    Put put = new Put(Bytes.toBytes(i));
                    put.addColumn(Bytes.toBytes("uf"), Bytes.toBytes("userid"), split[0].getBytes());
                    put.addColumn("uf".getBytes(), "friend".getBytes(),split[1].getBytes());
                    datas.add(put);
                }

                num = num + datas.size();
                System.out.println("---------num:" + num);
                if (datas.size() > 0){
                    ufTable.put(datas);
                }

                try {
                    Thread.sleep(10);
                } catch (InterruptedException e) {
                    throw new RuntimeException(e);
                }
            }
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    }
}

2.2普通模式写入hbase（buffer写入）

java 复制代码

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.*;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.kafka.clients.consumer.ConsumerConfig;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.common.serialization.StringDeserializer;

import java.io.IOException;
import java.time.Duration;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Properties;

/**
 * 将Kafka中的topic为userfriends中的数据消费到hbase中
 * hbase中的表为events_db:user_friend
 */
public class UserFriendToHB2 {
    static int num = 0; //计数器
    public static void main(String[] args) {
        Properties properties = new Properties();
        properties.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, "kb129:9092");
        properties.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
        properties.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG,StringDeserializer.class);
        properties.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG,"earliest");
        properties.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG,"false");
        properties.put(ConsumerConfig.GROUP_ID_CONFIG, "user_friend_group1");

        KafkaConsumer<String, String> consumer = new KafkaConsumer<>(properties);
        consumer.subscribe(Collections.singleton("userfriends"));

        //配置hbase信息，连接hbase数据库
        Configuration conf = HBaseConfiguration.create();
        conf.set(HConstants.HBASE_DIR, "hdfs://kb129:9000/hbase");
        conf.set(HConstants.ZOOKEEPER_QUORUM, "kb129");
        conf.set(HConstants.CLIENT_PORT_STR, "2181");

        Connection connection = null;
        try {
            connection = ConnectionFactory.createConnection(conf);
            BufferedMutatorParams bufferedMutatorParams = new BufferedMutatorParams(TableName.valueOf("events_db:user_friend"));
            bufferedMutatorParams.setWriteBufferPeriodicFlushTimeoutMs(10000);//设置超时flush时间最大值
            bufferedMutatorParams.writeBufferSize(10*1024*1024);//设置缓存大小flush
            BufferedMutator bufferedMutator = connection.getBufferedMutator(bufferedMutatorParams) ;

            ArrayList<Put> datas = new ArrayList<>();

            while (true){
                ConsumerRecords<String, String> poll = consumer.poll(Duration.ofMillis(100));
                datas.clear();  //每次for循环前清空datas
                for (ConsumerRecord<String, String> record : poll) {
                    //System.out.println(record.value());
                    String[] split = record.value().split(",");
                    int i = (split[0] + split[1]).hashCode();
                    Put put = new Put(Bytes.toBytes(i));
                    put.addColumn(Bytes.toBytes("uf"), Bytes.toBytes("userid"), split[0].getBytes());
                    put.addColumn("uf".getBytes(), "friend".getBytes(),split[1].getBytes());
                    datas.add(put);
                }

                num = num + datas.size();
                System.out.println("---------num:" + num);
                if (datas.size() > 0){
                    bufferedMutator.mutate(datas);
                }
            }
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    }
}

2.3设计模式写入hbase（buffer写入）

（1）Iworker接口

java 复制代码

public interface IWorker {
    void fillData(String targetName);
}

（2）worker实现类

java 复制代码

import nj.zb.kb23.kafkatohbase.oop.writer.IWriter;
import org.apache.kafka.clients.consumer.ConsumerConfig;
import org.apache.kafka.clients.consumer.ConsumerRecords;
import org.apache.kafka.clients.consumer.KafkaConsumer;
import org.apache.kafka.common.serialization.StringDeserializer;

import java.time.Duration;
import java.util.Collections;
import java.util.Properties;

public class Worker implements IWorker {
    private KafkaConsumer<String, String> consumer = null;
    private IWriter writer = null;

    public Worker(String topicName, String consumerGroupId, IWriter writer) {
        this.writer = writer;
        Properties properties = new Properties();
        properties.put(ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, "kb129:9092");
        properties.put(ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
        properties.put(ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class);
        properties.put(ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest");
        properties.put(ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, "false");
        properties.put(ConsumerConfig.GROUP_ID_CONFIG, consumerGroupId);

        consumer = new KafkaConsumer<>(properties);
        consumer.subscribe(Collections.singleton(topicName));
    }

    @Override
    public void fillData(String targetName) {
        int num = 0;
        while (true) {
            ConsumerRecords<String, String> records = consumer.poll(Duration.ofMillis(100));
            int returnNum = writer.write(targetName, records);
            num += returnNum;
            System.out.println("---------num:" + num);
        }
    }
}

（3）IWriter接口

java 复制代码

import org.apache.kafka.clients.consumer.ConsumerRecords;

/**
 * 完成kafka消费出的数据  ConsumerRecords 的组装和写入到指定类型的数据库 指定table 的工作
 */
public interface IWriter {
    int write(String targetTableName, ConsumerRecords<String, String> records);
}

（4）writer实现类

java 复制代码

import nj.zb.kb23.kafkatohbase.oop.handler.IParseRecord;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseConfiguration;
import org.apache.hadoop.hbase.HConstants;
import org.apache.hadoop.hbase.TableName;
import org.apache.hadoop.hbase.client.*;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.io.IOException;
import java.util.List;

public class HBaseWriter implements IWriter{
    private Connection connection = null;
    private BufferedMutator bufferedMutator = null;
    private IParseRecord handler = null;

    /**
     * 初始化HBaseWriter对象
     */
    public HBaseWriter(IParseRecord handler) {
        this.handler = handler;
        Configuration conf = HBaseConfiguration.create();
        conf.set(HConstants.HBASE_DIR, "hdfs://kb129:9000/hbase");
        conf.set(HConstants.ZOOKEEPER_QUORUM, "kb129");
        conf.set(HConstants.CLIENT_PORT_STR, "2181");

        try {
            connection = ConnectionFactory.createConnection(conf);
        } catch (IOException e) {
            throw new RuntimeException(e);
        }

    }

    private void getBufferedMutator(String targetTableName){
        BufferedMutatorParams bufferedMutatorParams = new BufferedMutatorParams(TableName.valueOf(targetTableName));
        bufferedMutatorParams.setWriteBufferPeriodicFlushTimeoutMs(10000);//设置超时flush时间最大值
        bufferedMutatorParams.writeBufferSize(10*1024*1024);//设置缓存大小flush

        if (bufferedMutator == null){
            try {
                bufferedMutator = connection.getBufferedMutator(bufferedMutatorParams);
            } catch (IOException e) {
                throw new RuntimeException(e);
            }
        }
    }


    @Override
    public int write(String targetTableName, ConsumerRecords<String, String> records) {
        if (records.count() > 0) {
            this.getBufferedMutator(targetTableName);
            List<Put> datas = handler.parse(records);
            try {
                bufferedMutator.mutate(datas);
            } catch (IOException e) {
                throw new RuntimeException(e);
            }
            return datas.size();
        }else {
            return 0;
        }
    }
}

（5）IParseRecord接口

java 复制代码

import org.apache.hadoop.hbase.client.Put;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.util.List;

/**
 * 将record 装配成 put
 */
public interface IParseRecord {
    List<Put> parse(ConsumerRecords<String, String> records);
}

（6）具体表对应的handler类（包装Put）

UsersHandler

java 复制代码

import org.apache.hadoop.hbase.client.Put;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.util.ArrayList;
import java.util.List;

public class UsersHandler implements IParseRecord{
    List<Put> datas = new ArrayList<>();
    @Override
    public List<Put> parse(ConsumerRecords<String, String> records) {
        datas.clear();
        for (ConsumerRecord<String, String> record : records) {
            String[] users = record.value().split(",");
            Put put = new Put(users[0].getBytes());
            put.addColumn("profile".getBytes(), "birthyear".getBytes(), users[2].getBytes());
            put.addColumn("profile".getBytes(), "gender".getBytes(), users[3].getBytes());
            put.addColumn("region".getBytes(), "locale".getBytes(), users[1].getBytes());
            if (users.length > 4){
                put.addColumn("registration".getBytes(), "joinedAt".getBytes(), users[4].getBytes());
            }
            if (users.length > 5){
                put.addColumn("region".getBytes(), "location".getBytes(), users[5].getBytes());
            }
            if (users.length > 6){
                put.addColumn("region".getBytes(), "timezone".getBytes(), users[6].getBytes());
            }
            datas.add(put);
        }
        return datas;
    }
}

TrainHandler

java 复制代码

import org.apache.hadoop.hbase.client.Put;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.util.ArrayList;
import java.util.List;

public class TrainHandler implements IParseRecord{
    List<Put> datas = new ArrayList<>();
    @Override
    public List<Put> parse(ConsumerRecords<String, String> records) {
        datas.clear();
        for (ConsumerRecord<String, String> record : records) {
            String[] trains = record.value().split(",");
            double random = Math.random();
            Put put = new Put((trains[0]+trains[1]+random).getBytes());
            put.addColumn("eu".getBytes(), "user".getBytes(), trains[0].getBytes());
            put.addColumn("eu".getBytes(), "event".getBytes(), trains[1].getBytes());
            put.addColumn("eu".getBytes(), "invited".getBytes(), trains[2].getBytes());
            put.addColumn("eu".getBytes(), "timestamp".getBytes(), trains[3].getBytes());
            put.addColumn("eu".getBytes(), "interested".getBytes(), trains[4].getBytes());
            put.addColumn("eu".getBytes(), "not_interested".getBytes(), trains[5].getBytes());
            datas.add(put);
        }
        return datas;
    }
}

EventsHandler

java 复制代码

import org.apache.hadoop.hbase.client.Put;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.util.ArrayList;
import java.util.List;

public class EventsHandler implements IParseRecord {
    List<Put> datas = new ArrayList<>();

    @Override
    public List<Put> parse(ConsumerRecords<String, String> records) {
        datas.clear();
        for (ConsumerRecord<String, String> record : records) {
            String[] events = record.value().split(",");
            Put put = new Put(events[0].getBytes());
            put.addColumn("creator".getBytes(), "user_id".getBytes(),events[1].getBytes());
            put.addColumn("schedule".getBytes(), "start_time".getBytes(),events[2].getBytes());
            put.addColumn("location".getBytes(), "city".getBytes(),events[3].getBytes());
            put.addColumn("location".getBytes(), "state".getBytes(),events[4].getBytes());
            put.addColumn("location".getBytes(), "zip".getBytes(),events[5].getBytes());
            put.addColumn("location".getBytes(), "country".getBytes(),events[6].getBytes());
            put.addColumn("location".getBytes(), "lat".getBytes(),events[7].getBytes());
            put.addColumn("location".getBytes(), "lng".getBytes(),events[8].getBytes());
            put.addColumn("remark".getBytes(), "common_words".getBytes(),events[9].getBytes());
            datas.add(put);
        }
        return datas;
    }
}

EventAttendHandler

java 复制代码

import org.apache.hadoop.hbase.client.Put;
import org.apache.hadoop.hbase.util.Bytes;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.clients.consumer.ConsumerRecords;

import java.util.ArrayList;
import java.util.List;

public class EventAttendHandler implements IParseRecord{
    @Override
    public List<Put> parse(ConsumerRecords<String, String> records) {
        List<Put> datas = new ArrayList<>();
        for (ConsumerRecord<String, String> record : records) {
            String[] splits = record.value().split(",");
            Put put = new Put((splits[0] + splits[1] + splits[2]).getBytes());
            put.addColumn(Bytes.toBytes("euat"), Bytes.toBytes("eventid"), splits[0].getBytes());
            put.addColumn("euat".getBytes(), "friendid".getBytes(),splits[1].getBytes());
            put.addColumn("euat".getBytes(), "state".getBytes(),splits[2].getBytes());
            datas.add(put);
        }
        return datas;
    }
}

（7）主程序

java 复制代码

import nj.zb.kb23.kafkatohbase.oop.handler.*;
import nj.zb.kb23.kafkatohbase.oop.worker.Worker;
import nj.zb.kb23.kafkatohbase.oop.writer.HBaseWriter;
import nj.zb.kb23.kafkatohbase.oop.writer.IWriter;
/**
 * 将Kafka中的topic为...中的数据消费到hbase中
 * hbase中的表为events_db:...
 */
public class KfkToHbTest {
    static int num = 0; //计数器

    public static void main(String[] args) {
        //IParseRecord handler = new EventAttendHandler();
        //IWriter writer = new HBaseWriter(handler);
        //String topic = "eventattendees";
        //String consumerGroupId = "eventattendees_group1";
        //String targetName = "events_db:event_attendee";
        //Worker worker = new Worker(topic, consumerGroupId, writer);
        //worker.fillData(targetName);

        /*EventsHandler eventsHandler = new EventsHandler();
        IWriter writer = new HBaseWriter(eventsHandler);
        Worker worker = new Worker("events", "events_group1", writer);
        worker.fillData("events_db:eventsb");*/

        /*UsersHandler usersHandler = new UsersHandler();
        IWriter writer = new HBaseWriter(usersHandler);
        Worker worker = new Worker("users_raw", "users_group1", writer);
        worker.fillData("events_db:users");*/

        TrainHandler trainHandler = new TrainHandler();
        IWriter writer = new HBaseWriter(trainHandler);
        Worker worker = new Worker("train", "train_group1", writer);
        worker.fillData("events_db:train2");

    }
}

3.HBase表映射至Hive中

sql 复制代码

create database if not exists events;
use events;

create external table hb_users(
    userId string,
    birthyear int,
    gender string,
    locale string,
    location string,
    timezone string,
    joinedAt string
)stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
with SERDEPROPERTIES (
    'hbase.columns.mapping'=':key,profile:birthyear,profile:gender,region:locale,region:location,region:timezone,registration:joinedAt'
    )
tblproperties ('hbase.table.name'='events_db:users');

select * from hb_users limit 3;
select count(1) from hb_users;

--orc格式创建内部表存储映射外部表，安全保存数据，创建好可以直接删除hbase中的表
create table users stored as orc as select * from hb_users;
select * from users limit 3;
select count(1) from users;
drop table hb_users;

--38209  1494
select count(*) from users where birthyear is null;

select round(avg(birthyear), 0) from users;
select `floor`(avg(birthyear)) from users;

-- 处理空字段，覆盖写入
with
    tb as ( select `floor`(avg(birthyear)) avgAge from users ),
    tb2 as ( select userId, nvl(birthyear, tb.avgAge),gender,locale,location,timezone,joinedAt from users,tb)
insert overwrite table users
select * from tb2;

-- 查询到性别中空字符串109个
select count(gender) count from users where gender is null or gender = "";

--------------------------------------------------------
create external table hb_events(
    event_id string,
    user_id string,
    start_time string,
    city string,
    state string,
    zip string,
    country string,
    lat float,
    lng float,
    common_words string
)stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
with SERDEPROPERTIES (
    'hbase.columns.mapping'=':key,creator:user_id,schedule:start_time,location:city,location:state,location:zip,location:country,location:lat,location:lng,remark:common_words'
    )
tblproperties ('hbase.table.name'='events_db:events');

select * from hb_events limit 10;
create table events stored as orc as select * from hb_events;
select count(*) from hb_events;
select count(*) from events;
drop table hb_events;

select event_id from events group by event_id having count(event_id) >1;
with
    tb as (select event_id, row_number() over (partition by event_id) rn from events)
select event_id from tb where rn > 1;

select user_id, count(event_id) num from events group by user_id order by num desc;

-----------------------------------------------------
create external table if not exists hb_user_friend(
    row_key string,
    userid string,
    friendid string
)stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
with SERDEPROPERTIES ('hbase.columns.mapping'=':key,uf:userid,uf:friend')
tblproperties ('hbase.table.name'='events_db:user_friend');


select * from hb_user_friend limit 3;
create table user_friend stored as orc as select * from hb_user_friend;
select count(*) from hb_user_friend;
select count(*) from user_friend;
drop table hb_user_friend;

-----------------------------------------------------------
create external table if not exists hb_event_attendee(
    row_key string,
    eventid string,
    friendid string,
    attendtype string
)stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
with SERDEPROPERTIES ('hbase.columns.mapping'=':key,euat:eventid,euat:friendid,euat:state')
tblproperties ('hbase.table.name'='events_db:event_attendee');

select * from hb_event_attendee limit 3;
select count(*) from hb_event_attendee;
create table event_attendee stored as orc as select * from hb_event_attendee;
select * from event_attendee limit 3;
select count(*) from event_attendee;
drop table hb_event_attendee;


--------------------------------------------------------------
create external table if not exists hb_train(
    row_key string,
    userid string,
    eventid string,
    invited string,
    `timestamp` string,
    interested string
)stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
with SERDEPROPERTIES ('hbase.columns.mapping'=':key,eu:user,eu:event,eu:invited,eu:timestamp,eu:interested')
tblproperties ('hbase.table.name'='events_db:train');

select * from hb_train limit 3;
select count(*) from hb_train;
create table train stored as orc as select * from hb_train;
select * from train limit 3;
select count(*) from train;
drop table hb_train;

-----------------------------------------------
create external table locale(
    locale_id int,
    locale string
)
row format delimited fields terminated by '\t'
location '/events/data/locale';

select * from locale;

create external table time_zone(
    time_zone_id int,
    time_zone string
)
row format delimited fields terminated by ','
location '/events/data/timezone';

select * from time_zone;