Alex 的 Hadoop 菜鸟教程: 第21课 不只是在HBase中用SQL：Phoenix
声明
- 本文基于 Centos6.x + CDH5.x
什么是Phoenix
Phoenix的团队用了一句话概括Phoenix："We put the SQL back in NoSQL" 意思是：我们把SQL又放回NoSQL去了！这边说的NoSQL专指HBase，意思是可以用SQL语句来查询Hbase，你可能会说：“Hive和Impala也可以啊！”。但是Hive和Impala还可以查询文本文件，Phoenix的特点就是，它只能查Hbase，别的类型都不支持！但是也因为这种专一的态度，让Phoenix在Hbase上查询的性能超过了Hive和Impala！
安装Phoenix
之前的组件都是通过CDH来安装的，但是这回就跟Cloudera完全没关系了。从 Apache Phoenix Download 下载Phoenix包，不过这个镜像有点慢，我把包上传到CSDN了，下载地址见下面的版本对应
Phoenix跟Hbase的版本对应
- Phoenix 2.x - HBase 0.94.x
- Phoenix 3.x - HBase 0.94.x 下载地址
- Phoenix 4.x - HBase 0.98.1+ 下载地址 官网下载地址
这里我用4.2.2，下载好后，解压开，把 phoenix-4.2.2-server.jar 拷贝到所有RegionServer的lib目录下 /usr/lib/hbase/lib
cp phoenix-4.2.2-server.jar /usr/lib/hbase/lib
然后重启所有regionserver
service hbase-regionserver restart
使用Phoenix
把 phoenix-4.2.2-bin.tar.gz 解压出来的 phoenix-4.2.2-bin 文件夹也上传到host1上，然后到bin目录下执行(其实在本机也可以，只是我本机没有Python环境而Centos天生有Python)
如果是windows下下载的，得先在centos上给bin文件夹里面的.py文件赋上执行权限
[root@host1 ~]# cd phoenix-4.2.2-bin/ [root@host1 phoenix-4.2.2-bin]# cd bin [root@host1 bin]# chmod +x *.py
- 批处理方式
- 命令行方式
- GUI方式
- JDBC调用
批处理方式
我们建立sql 名叫 us_population.sql 内容是
CREATE TABLE IF NOT EXISTS us_population ( state CHAR(2) NOT NULL, city VARCHAR NOT NULL, population BIGINT CONSTRAINT my_pk PRIMARY KEY (state, city));
建立一个文件 us_population.csv
NY,New York,8143197 CA,Los Angeles,3844829 IL,Chicago,2842518 TX,Houston,2016582 PA,Philadelphia,1463281 AZ,Phoenix,1461575 TX,San Antonio,1256509 CA,San Diego,1255540 TX,Dallas,1213825 CA,San Jose,912332
再创建一个文件 us_population_queries.sql
SELECT state as "State",count(city) as "City Count",sum(population) as "Population Sum" FROM us_population GROUP BY state ORDER BY sum(population) DESC;
然后一起执行
phoenix-4.2.2-bin/bin/psql.py host1,host2:2181 us_population.sql us_population.csv us_population_queries.sql
这边记得把 host1 和 host2 换成你的zookeeper地址
这条命令你同时做了 创建一个表，插入数据，查询结果 三件事情
[root@host1 ~]# phoenix-4.2.2-bin/bin/psql.py host1,host2:2181 us_population.sql us_population.csv us_population_queries.sql 15/03/04 17:14:23 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable 15/03/04 17:14:24 WARN impl.MetricsConfig: Cannot locate configuration: tried hadoop-metrics2-phoenix.properties,hadoop-metrics2.properties no rows upserted Time: 0.726 sec(s) csv columns from database. CSV Upsert complete. 10 rows upserted Time: 0.103 sec(s) St City Count Population Sum -- ---------------------------------------- ---------------------------------------- NY 1 8143197 CA 3 6012701 TX 3 4486916 IL 1 2842518 PA 1 1463281 AZ 1 1461575 Time: 0.048 sec(s)
用hbase shell 看下会发现多出来一个 US_POPULATION 表，用scan 命令查看一下这个表的数据
hbase(main):002:0> scan 'US_POPULATION' ROW COLUMN+CELL AZPhoenix column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x16MG AZPhoenix column=0:_0, timestamp=1425460467206, value= CALos Angeles column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00:\xAA\xDD CALos Angeles column=0:_0, timestamp=1425460467206, value= CASan Diego column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x13(t CASan Diego column=0:_0, timestamp=1425460467206, value= CASan Jose column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x0D\xEB\xCC CASan Jose column=0:_0, timestamp=1425460467206, value= ILChicago column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00+_\x96 ILChicago column=0:_0, timestamp=1425460467206, value= NYNew York column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00|A] NYNew York column=0:_0, timestamp=1425460467206, value= PAPhiladelphia column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x16S\xF1 PAPhiladelphia column=0:_0, timestamp=1425460467206, value= TXDallas column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x12\x85\x81 TXDallas column=0:_0, timestamp=1425460467206, value= TXHouston column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x1E\xC5F TXHouston column=0:_0, timestamp=1425460467206, value= TXSan Antonio column=0:POPULATION, timestamp=1425460467206, value=\x80\x00\x00\x00\x00\x13,= TXSan Antonio column=0:_0, timestamp=1425460467206, value= 10 row(s) in 0.2220 seconds
会发现
- 之前定义的PRIMARY KEY 为 state, city ，于是Phoenix就把你输入的state 和 city的值拼起来成为rowkey
- 其他的字段还是按照列名去保存，默认的列簇为 0
- 还有一个0:_0 这个列是没有值的，这个是Phoenix处于性能方面考虑增加的一个列，不用管这个列
命令行方式
然后执行sqlline.py
$ ./sqlline.py localhost
可以进入命令行模式
0: jdbc:phoenix:localhost>
退出命令行的方式是执行 !quit
0: jdbc:phoenix:localhost>!quit命令开头需要一个感叹号，使用help可以打印出所有命令
0: jdbc:phoenix:localhost> help
!all Execute the specified SQL against all the current connections
!autocommit Set autocommit mode on or off
!batch Start or execute a batch of statements
!brief Set verbose mode off
!call Execute a callable statement
!close Close the current connection to the database
!closeall Close all current open connections
!columns List all the columns for the specified table
!commit Commit the current transaction (if autocommit is off)
!connect Open a new connection to the database.
!dbinfo Give metadata information about the database
!describe Describe a table
!dropall Drop all tables in the current database
!exportedkeys List all the exported keys for the specified table
!go Select the current connection
!help Print a summary of command usage
!history Display the command history
!importedkeys List all the imported keys for the specified table
!indexes List all the indexes for the specified table
!isolation Set the transaction isolation for this connection
!list List the current connections
!manual Display the SQLLine manual
!metadata Obtain metadata information
!nativesql Show the native SQL for the specified statement
!outputformat Set the output format for displaying results
(table,vertical,csv,tsv,xmlattrs,xmlelements)
!primarykeys List all the primary keys for the specified table
!procedures List all the procedures
!properties Connect to the database specified in the properties file(s)
!quit Exits the program
!reconnect Reconnect to the database
!record Record all output to the specified file
!rehash Fetch table and column names for command completion
!rollback Roll back the current transaction (if autocommit is off)
!run Run a script from the specified file
!save Save the current variabes and aliases
!scan Scan for installed JDBC drivers
!script Start saving a script to a file
!set Set a sqlline variable
!sql Execute a SQL command
!tables List all the tables in the database
!typeinfo Display the type map for the current connection
!verbose Set verbose mode on
建立employee的映射表
数据准备
然后我们来建立一个映射表，映射我之前建立过的一个hbase表 employeehbase(main):003:0> describe 'employee'
DESCRIPTION ENABLED
'employee', {NAME => 'company', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLICATION_SCOPE => '0', VERSIONS => true
'1', COMPRESSION => 'NONE', MIN_VERSIONS => '0', TTL => 'FOREVER', KEEP_DELETED_CELLS => 'false', BLOCKSIZE => '65536', I
N_MEMORY => 'false', BLOCKCACHE => 'true'}, {NAME => 'family', DATA_BLOCK_ENCODING => 'NONE', BLOOMFILTER => 'ROW', REPLIC
ATION_SCOPE => '0', VERSIONS => '1', COMPRESSION => 'NONE', MIN_VERSIONS => '0', TTL => 'FOREVER', KEEP_DELETED_CELLS => '
false', BLOCKSIZE => '65536', IN_MEMORY => 'false', BLOCKCACHE => 'true'}
1 row(s) in 0.1120 seconds
可以看出employee有连个列簇 company 和 family
hbase(main):016:0> scan 'employee' ROW COLUMN+CELL row1 column=company:name, timestamp=1425537923391, value=ted row1 column=company:position, timestamp=1425537950471, value=worker row1 column=family:tel, timestamp=1425537956413, value=13600912345 row2 column=family:tel, timestamp=1425537994087, value=18942245698 row2 column=family:name, timestamp=1425537975610, value=michael row2 column=family:position, timestamp=1425537985594, value=manager 2 row(s) in 0.0340 seconds
有两条数据。如果没有这些数据的同学可以用以下命令创建
create 'employee','company','family' put 'employee','row1','company:name','ted' put 'employee','row1','company:position','worker' put 'employee','row1','family:tel','13600912345' put 'employee','row2','company:name','michael' put 'employee','row2','company:position','manager' put 'employee','row2','family:tel','1894225698' scan 'employee'
关于映射表
在建立映射表之前要说明的是，Phoenix是大小写敏感的，并且所有命令都是大写，如果你建的表名没有用双引号括起来，那么无论你输入的是大写还是小写，建立出来的表名都是大写的，如果你需要建立出同时包含大写和小写的表名和字段名，请把表名或者字段名用双引号括起来
你可以建立读写的表或者只读的表，他们的区别如下
- 读写表：如果你定义的列簇不存在，会被自动建立出来，并且赋以空值
- 只读表：你定义的列簇必须事先存在
建立映射
0: jdbc:phoenix:localhost> CREATE TABLE IF NOT EXISTS "employee" ("no" CHAR(4) NOT NULL PRIMARY KEY, "company"."name" VARCHAR(30),"company"."position" VARCHAR(20), "family"."tel" CHAR(11), "family"."age" INTEGER);
2 rows affected (1.745 seconds)
这行语句有几个注意点
- IF NOT EXISTS可以保证如果已经有建立过这个表，配置不会被覆盖
- 作为rowkey的字段用 PRIMARY KEY标定
- 列簇用 columnFamily.columnName 来表示
- family.age 是新增的字段，我之前建立测试数据的时候没有建立这个字段的原因是在hbase shell下无法直接写入数字型，等等我用UPSERT 命令插入数据的时候你就可以看到真正的数字型在hbase 下是如何显示的
建立好后，查询一下数据
0: jdbc:phoenix:localhost> SELECT * FROM "employee"; +------+--------------------------------+----------------------+-------------+------------------------------------------+ | no | name | position | tel | age | +------+--------------------------------+----------------------+-------------+------------------------------------------+ | row1 | ted | worker | 13600912345 | null | | row2 | michael | manager | 1894225698 | null | +------+--------------------------------+----------------------+-------------+------------------------------------------+
插入/更改数据
插入或者更改数据在Phoenix里面是一个命令叫 UPSERT 意思是 update + insert
我们插入一条数据试试
UPSERT INTO "employee" VALUES ('row3','billy','worker','16974681345',33);
查询一下数据
0: jdbc:phoenix:localhost> SELECT * FROM "employee"; +------+--------------------------------+----------------------+-------------+------------------------------------------+ | no | name | position | tel | age | +------+--------------------------------+----------------------+-------------+------------------------------------------+ | row1 | ted | worker | 13600912345 | null | | row2 | michael | manager | 1894225698 | null | | row3 | billy | worker | 16974681345 | 33 | +------+--------------------------------+----------------------+-------------+------------------------------------------+ 3 rows selected (0.195 seconds)
我们去hbase里面看一下数据
hbase(main):054:0> scan 'employee' ROW COLUMN+CELL row1 column=company:_0, timestamp=1425543735420, value= row1 column=company:name, timestamp=1425543735274, value=ted row1 column=company:position, timestamp=1425543735323, value=worker row1 column=family:tel, timestamp=1425543735420, value=13600912345 row2 column=company:_0, timestamp=1425543735767, value= row2 column=company:name, timestamp=1425543735608, value=michael row2 column=company:position, timestamp=1425543735720, value=manager row2 column=family:tel, timestamp=1425543735767, value=1894225698 row3 column=company:_0, timestamp=1425543857594, value= row3 column=company:name, timestamp=1425543857594, value=billy row3 column=company:position, timestamp=1425543857594, value=worker row3 column=family:age, timestamp=1425543857594, value=\x80\x00\x00! row3 column=family:tel, timestamp=1425543857594, value=16974681345 3 row(s) in 0.0650 seconds
TABLE 跟 VIEW 的区别
建立VIEW的语句跟TABLE一样，就是把TABLE关键字换成VIEW而已，但是他们有本质的区别:
- 如果你有一张Hbase的表，并且以前已经有数据，最好建立VIEW，因为一旦建立了TABLE, Phoenix 会认为这张table完全是属于Phoenix的，就算你只是删除Phoenix里面的表，Hbase里面对应的表也会一起删除掉
- 如果你之前没有这张Hbase的表，你想建立全新的Phoenix表，就用Table，但是要记住所有的操作都要经过Phoenix，那张同时被建立的表，只是一张附属表，不要试图往里面手动维护数据，忘记它的存在
用GUI方式
GUI方式的安装方法在 http://phoenix.apache.org/installation.html 这边不讲了，因为我自己也没搞起来，而且那个界面实在太丑了，看了不忍心使用。
JDBC调用
打开Eclipse建立一个简单的Maven项目 play-phoenix
pom.xml的内容是
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"> <modelVersion>4.0.0</modelVersion> <groupId>org.crazycake</groupId> <artifactId>play-phoenix</artifactId> <version>0.0.1-SNAPSHOT</version> <packaging>jar</packaging> <name>play-phoenix</name> <url>http://maven.apache.org</url> <properties> <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding> </properties> <repositories> <repository> <id>apache release</id> <url>https://repository.apache.org/content/repositories/releases/</url> </repository> </repositories> <build> <plugins> <plugin> <artifactId>maven-compiler-plugin</artifactId> <version>2.0.2</version> <configuration> <source>1.7</source> <target>1.7</target> <encoding>UTF-8</encoding> <optimise>true</optimise> <compilerArgument>-nowarn</compilerArgument> </configuration> </plugin> <plugin> <groupId>org.apache.maven.plugins</groupId> <artifactId>maven-shade-plugin</artifactId> <version>2.3</version> <configuration> <transformers> <transformer implementation="org.apache.maven.plugins.shade.resource.ApacheLicenseResourceTransformer"> </transformer> </transformers> </configuration> <executions> <execution> <phase>package</phase> <goals> <goal>shade</goal> </goals> </execution> </executions> </plugin> </plugins> </build> <dependencies> <dependency> <groupId>junit</groupId> <artifactId>junit</artifactId> <version>3.8.1</version> <scope>test</scope> </dependency> <dependency> <groupId>org.apache.phoenix</groupId> <artifactId>phoenix-core</artifactId> <version>4.2.2</version> </dependency> </dependencies> </project>
phoenix 4.2.2 使用jdk1.7编译的，如果你只有1.6就用 4.2.0
<dependency> <groupId>org.apache.phoenix</groupId> <artifactId>phoenix-core</artifactId> <version>4.2.0</version> </dependency>
我们建立一个类 PhoenixManager
package org.crazycake.play_phoenix;
import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.ResultSet;
import java.sql.SQLException;
import java.sql.Statement;
public class PhoenixManager {
public static void main(String[] args) throws SQLException {
Connection conn = null;
Statement stat = null;
ResultSet rs = null;
try {
Class.forName("org.apache.phoenix.jdbc.PhoenixDriver");
conn = DriverManager.getConnection("jdbc:phoenix:host1,host2:2181");
stat = conn.createStatement();
rs = stat.executeQuery("select * from \"employee\"");
while(rs.next()){
System.out.println("no: " + rs.getString("no"));
System.out.println("name: " + rs.getString("name"));
System.out.println("position: " + rs.getString("position"));
System.out.println("age: " + rs.getInt("age"));
}
} catch (Throwable e) {
e.printStackTrace();
} finally{
if(rs != null){
rs.close();
}
if(stat != null){
stat.close();
}
if(conn != null){
conn.close();
}
}
}
}
运行下，结果为
no: row1 name: ted position: worker age: 0 no: row2 name: michael position: manager age: 0 no: row3 name: billy position: worker age: 33
搞定！
结语
至此所有Hadoop必学的组件已经完成，菜鸟课程已经完成！后续的非必学组件我就看心情更新了！ :-)
参考资料
- http://phoenix.apache.org/Phoenix-in-15-minutes-or-less.html