Flink构造宽表实时入库案例介绍

2024-01-11 13:00:05
开发
56

1. 安装包准备

Flink 1.15.4 安装包

Flink cdc的mysql连接器

Flink sql的sdb连接器

MySQL驱动

SDB驱动

Flink jdbc的mysql连接器

2. 入库流程图

3. Flink安装部署

上传Flink压缩包到服务器，并解压

tar -zxvf flink-1.14.5-bin-scala_2.11.tgz -C /opt/

复制依赖至Flink中

cp sdb-flink-connector-3.4.8-jar-with-dependencies.jar /opt/flink-1.14.5/lib
cp sequoiadb-driver-3.4.8.jre8.jar /opt/flink-1.14.5/lib
cp flink-sql-connector-mysql-cdc-2.2.1.jar /opt/flink-1.14.5/lib
cp flink-connector-jdbc_2.11-1.14.6.jar /opt/flink-1.14.5/lib

修改flink-conf.yaml文件

vi conf/flink-conf.yaml

### 配置Master的机器名（IP地址）
jobmanager.rpc.address: sdb1
### 配置每个taskmanager 生成的临时文件夹
io.tmp.dirs: /opt/flink-1.14.5/tmp

修改master文件

vi conf/masters

#作为master的ip和端口号
upgrade1:8081

修改worker文件

vi conf/workers

#集群主机名
upgrade1
upgrade2
upgrade3

拷贝到集群其他机器

scp -r /opt/flink-1.14.5 sdbadmin@upgrade2:/opt/
scp -r /opt/flink-1.14.5 sdbadmin@upgrade3:/opt/

启动flink集群

[sdbadmin@upgrade1 flink-1.14.5]$ ./bin/start-cluster.sh

启动flink-SQL

[sdbadmin@upgrade1 flink-1.14.5]$ ./bin/sql-client.sh

4. 实时入库

编写造数程序进行造数

4.1 环境准备

4.1.1 开启mysql的binlog

创建binlog文件夹

[sdbadmin@upgrade1 mysql]$ mkdir /opt/sequoiasql/mysql/database/3306/binlog

开启binlog

vim /opt/sequoiasql/mysql/database/3306/auto.cnf

>>配置以下内容：
log_bin=/opt/sequoiasql/mysql/database/3306/binlog
binlog_format=ROW
expire_logs_days=1
server_id=1

配置完成之后，重启mysql

[sdbadmin@upgrade1 mysql]$ ./bin/sdb_mysql_ctl stop myinst
[sdbadmin@upgrade1 mysql]$ ./bin/sdb_mysql_ctl start myinst

4.1.2 创建mysql表

创建库

create database sbtest；
use sbtest；

创建表

CREATE TABLE sbtest1 (
    id INT UNSIGNED AUTO_INCREMENT,
    uuid INT(10),
    name1 CHAR(120),
    age INT(4),
    time1 DATETIME,
    PRIMARY KEY(id)
);

CREATE TABLE sbtest2 (
    id INT UNSIGNED AUTO_INCREMENT,
    uuid INT(10),
    name2 CHAR(120),
    age INT(4),
    time1 DATETIME,
    PRIMARY KEY(id)
);

CREATE TABLE sbtest3 (
    id INT UNSIGNED AUTO_INCREMENT,
    uuid INT(10),
    name3 CHAR(120),
    age INT(4),
    time1 DATETIME,
    PRIMARY KEY(id)
);

创建flink入库表

CREATE TABLE sbtest4 (
    id INT UNSIGNED AUTO_INCREMENT,
    uuid INT(10),
    name1 CHAR(120),
    name2 CHAR(120),
    name3 CHAR(120),
    age INT(4),
    time1 DATETIME,
    PRIMARY KEY(id)
);

4.1.3 创建flink映射表

需要用到flink-sql-connector-mysql-cdc-2.2.1.jar

CREATE TABLE sbtest1_mysql (
    id INT,
    uuid INT,
    name1 CHAR(120),
    age INT,
    time1 TIMESTAMP,
    PRIMARY KEY (id) NOT ENFORCED
  ) WITH (
    'connector' = 'mysql-cdc',
    'hostname' = '192.168.223.135',
    'port' = '3306',
    'username' = 'root',
    'password' = 'root',
    'database-name' = 'sbtest',
    'table-name' = 'sbtest1'
);

CREATE TABLE sbtest2_mysql (
    id INT,
    uuid INT,
    name2 CHAR(120),
    age INT,
    time1 TIMESTAMP,
    PRIMARY KEY (id) NOT ENFORCED
  ) WITH (
    'connector' = 'mysql-cdc',
    'hostname' = '192.168.223.135',
    'port' = '3306',
    'username' = 'root',
    'password' = 'root',
    'database-name' = 'sbtest',
    'table-name' = 'sbtest2'
);

CREATE TABLE sbtest3_mysql (
    id INT,
    uuid INT,
    name3 CHAR(120),
    age INT,
    time1 TIMESTAMP,
    PRIMARY KEY (id) NOT ENFORCED
  ) WITH (
    'connector' = 'mysql-cdc',
    'hostname' = '192.168.223.135',
    'port' = '3306',
    'username' = 'root',
    'password' = 'root',
    'database-name' = 'sbtest',
    'table-name' = 'sbtest3'
);

创建flink --> mysql入库映射表

需要用到flink-connector-jdbc_2.11-1.14.6.jar

CREATE TABLE sbtest4_mysql (
    id BIGINT,
    uuid INT,
    name1 CHAR(120),
    name2 CHAR(120),
    name3 CHAR(120),
    age INT,
    time1 TIMESTAMP,
    PRIMARY KEY (id) NOT ENFORCED
  ) WITH (
    'connector' = 'jdbc',
    'url' = 'jdbc:mysql://192.168.223.135:3306/sbtest',
    'username' = 'root',
    'password' = 'root',
    'table-name' = 'sbtest4'
);

创建flink --> mysql入库映射表

需要用到sdb-flink-connector-3.4.8-jar-with-dependencies.jar

CREATE TABLE sbtest_sdb (
    id BIGINT,
    uuid INT,
    name1 CHAR(120),
    name2 CHAR(120),
    name3 CHAR(120),
    age INT,
    time1 TIMESTAMP,
    PRIMARY KEY (id) NOT ENFORCED
) WITH (
    'connector' = 'sequoiadb',
    'bulksize' = '1',
    'hosts' = '192.168.223.135:11810',
    'collectionspace' = 'sbtest',
    'collection' = 'sbtest4'
);

4.2 MySQL实时入库

4.2.1 Flink left join

select sdb1.id, sdb1.uuid, sdb1.name1, sdb2.name2, sdb3.name3, sdb1.age, sdb1.time1
from sbtest1_mysql sdb1
left join sbtest2_mysql sdb2
on sdb1.id = sdb2.id
left join sbtest3_mysql sdb3
on sdb1.id = sdb3.id;

4.2.2 mysql实时入库

insert into sbtest4_mysql select sdb1.id, sdb1.uuid, sdb1.name1, sdb2.name2, sdb3.name3, sdb1.age, sdb1.time1
from sbtest1_mysql sdb1
left join sbtest2_mysql sdb2
on sdb1.id = sdb2.id
left join sbtest3_mysql sdb3
on sdb1.id = sdb3.id;

查看Flink任务

查看可以成功入库

4.3 SDB实时入库

4.3.1 Flink left join

4.3.2 sdb实时入库

insert into sbtest_sdb select sdb1.id, sdb1.uuid, sdb1.name1, sdb2.name2, sdb3.name3, sdb1.age, sdb1.time1
from sbtest1_mysql sdb1
left join sbtest2_mysql sdb2
on sdb1.id = sdb2.id
left join sbtest3_mysql sdb3
on sdb1.id = sdb3.id;

查看Flink任务

显示已经成功入库

原文地址:https://blog.csdn.net/u014439239/article/details/135495181 本文来自互联网用户投稿，该文观点仅代表作者本人，不代表本站立场。本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如若转载，请注明出处：https://www.suanlizi.com/kf/1745309580910530560.html 如若内容造成侵权/违法违规/事实不符，请联系《酸梨子》网邮箱：1419361763@qq.com进行投诉反馈，一经查实，立即删除！

阅读全部

Flink构造宽表实时入库案例介绍

相关推荐

最近更新

热门阅读