基于FLINK + PAIMON + StarRocks 分层构建流式湖仓

j结合Paimon与StarRocks构建流式湖仓-实时计算 Flink版-阿里云

参考：https://blog.csdn.net/MDZ_1122333322/article/details/145651839

FLINK1.20 + FLINK CDC + MINIO3 + PAIMON1.2 + StarRocks3.5.7 + JDK11 + streampark2.1

Mysql SQL FLINK CDC 同步到 PAIMON (基于 minio)

在streampark上 创建 dwd到dws层的Flink任务

通过starrocks查询PAIMON的数据，并构建异步物化视图

在mysql中创建名称为order_dw的数据库

create database order_dw;

建表并写入数据

CREATE TABLE `orders` (
  order_id bigint not null primary key,
  user_id varchar(50) not null,
  shop_id bigint not null,
  product_id bigint not null,
  buy_fee bigint not null,   
  create_time timestamp not null,
  update_time timestamp not null default now(),
  state int not null
);

CREATE TABLE `orders_pay` (
  pay_id bigint not null primary key,
  order_id bigint not null,
  pay_platform int not null, 
  create_time timestamp not null
);

CREATE TABLE `product_catalog` (
  product_id bigint not null primary key,
  catalog_name varchar(50) not null
);

-- 准备数据
INSERT INTO product_catalog VALUES(1, 'phone_aaa'),(2, 'phone_bbb'),(3, 'phone_ccc'),(4, 'phone_ddd'),(5, 'phone_eee');

INSERT INTO orders VALUES
(100001, 'user_001', 12345, 1, 5000, '2023-02-15 16:40:56', '2023-02-15 18:42:56', 1),
(100002, 'user_002', 12346, 2, 4000, '2023-02-15 15:40:56', '2023-02-15 18:42:56', 1),
(100003, 'user_003', 12347, 3, 3000, '2023-02-15 14:40:56', '2023-02-15 18:42:56', 1),
(100004, 'user_001', 12347, 4, 2000, '2023-02-15 13:40:56', '2023-02-15 18:42:56', 1),
(100005, 'user_002', 12348, 5, 1000, '2023-02-15 12:40:56', '2023-02-15 18:42:56', 1),
(100006, 'user_001', 12348, 1, 1000, '2023-02-15 11:40:56', '2023-02-15 18:42:56', 1),
(100007, 'user_003', 12347, 4, 2000, '2023-02-15 10:40:56', '2023-02-15 18:42:56', 1);

INSERT INTO orders_pay VALUES
(2001, 100001, 1, '2023-02-15 17:40:56'),
(2002, 100002, 1, '2023-02-15 17:40:56'),
(2003, 100003, 0, '2023-02-15 17:40:56'),
(2004, 100004, 0, '2023-02-15 17:40:56'),
(2005, 100005, 0, '2023-02-15 18:40:56'),
(2006, 100006, 0, '2023-02-15 18:40:56'),
(2007, 100007, 0, '2023-02-15 18:40:56');

下载：

Central Repository: org/apache/paimon/paimon-flink-action/1.2.0

wget https://repo1.maven.org/maven2/org/apache/paimon/paimon-flink-action/1.2.0/paimon-flink-action-1.2.0.jar .

下载 flink-sql-connector-mysql-cdc-3.1.x.jar

Maven Repository: org.apache.flink >> flink-sql-connector-mysql-cdc >> 3.1.1

增加的主要依赖包如下：

copy到flink/lib下 重启flink集群

参考：Mysql CDC | Apache Paimon

FLINK CDC 整体 同到 PAIMON，采用 paimon-flink-action-1.2.0.jar

方式一 在flink-1.20.0所在服务期 直接提供 flink 命令

# 提交 Flink 作业
/opt/flink-1.20.0/bin/flink run \
  -Dexecution.checkpointing.interval=10s \
  -Dstate.checkpoints.dir=file:///tmp/flink-checkpoints \
  /opt/flink-1.20.0/lib/paimon-flink-action-1.2.0.jar \
  mysql_sync_database \
  --warehouse s3://warehouse/wh \
  --database order_dw \
  --mysql_conf hostname=192.168.1.247 \
  --mysql_conf port=3306 \
  --mysql_conf username=root \
  --mysql_conf password=123456 \
  --mysql_conf database-name=order_dw \
  --catalog_conf s3.endpoint=http://192.168.1.243:9000 \
  --catalog_conf s3.access-key=minio \
  --catalog_conf s3.secret-key=minio@123 \
  --catalog_conf s3.region=us-east-1 \
  --catalog_conf s3.path.style.access=true \
  --table_conf bucket=1 \
  --table_conf changelog-producer=input \
  --table_conf sink.parallelism=1

方式二 ：在streampark上提交任务

参考:在streampark运行paimon-flink-action-1.20.0.jar-CSDN博客

1、上传相关jar包到streampark所在服务的 /opt/flink-1.20.0/lib

rm -rf flink-sql-connector-mysql-cdc-3.0.1.jar （要用3.1）

2、在streampark上配置任务

上传

添加程序参数

mysql_sync_database
--warehouse s3://warehouse/wh
--database order_dw
--mysql_conf hostname=192.168.1.247
--mysql_conf port=3306
--mysql_conf username=root
--mysql_conf password=123456
--mysql_conf database-name=order_dw
--catalog_conf s3.endpoint=http://192.168.1.243:9000
--catalog_conf s3.access-key=minio
--catalog_conf s3.secret-key=minio@123
--catalog_conf s3.region=us-east-1
--catalog_conf s3.path.style.access=true
--table_conf bucket=1
--table_conf changelog-producer=input
--table_conf sink.parallelism=1

运行成功：

---

验证数据：

数据已经同步过来了

starrocks表已经有了

---

构建 DWD ：dwd_orders

-- Step 2: ?? Paimon Catalog
CREATE CATALOG paimon_minio WITH (
  'type' = 'paimon',
  'warehouse' = 's3://warehouse/wh',
  's3.endpoint' = 'http://192.168.1.243:9000',
  's3.access-key' = 'minio',
  's3.secret-key' = 'minio@123',
  's3.region' = 'us-east-1',
  's3.path.style.access' = 'true'
);
-- ? ??? Paimon Catalog?????
USE CATALOG paimon_minio;
CREATE TABLE IF NOT EXISTS paimon_minio.order_dw.dwd_orders (
  order_id BIGINT,
  order_user_id STRING,
  order_shop_id BIGINT,
  order_product_id BIGINT,
  order_product_catalog_name STRING,
  order_fee BIGINT,
  order_create_time TIMESTAMP,
  order_update_time TIMESTAMP,
  order_state INT,
  pay_id BIGINT,
  pay_platform INT COMMENT 'platform 0: phone, 1: pc',
  pay_create_time TIMESTAMP,
  PRIMARY KEY (order_id) NOT ENFORCED
) WITH (
  'merge-engine' = 'partial-update',
  -- ????????????????
  'partial-update.romove-record-on-delete' = 'true',
  -- ????
  'changelog-producer' = 'lookup' -- ??lookup??????????????????
);
SET
  'execution.checkpointing.max-concurrent-checkpoints' = '3';
SET
  'table.exec.sink.upsert-materialize' = 'NONE';
SET
  'execution.checkpointing.interval' = '10s';
SET
  'execution.checkpointing.min-pause' = '10s';
-- Paimon?????????????????INSERT???????????????UNION ALL?
INSERT INTO
  paimon_minio.order_dw.dwd_orders
SELECT
  o.order_id,
  o.user_id,
  o.shop_id,
  o.product_id,
  dim.catalog_name,
  o.buy_fee,
  o.create_time,
  o.update_time,
  o.state,
  CAST(NULL AS BIGINT) AS pay_id,
  CAST(NULL AS INT) AS pay_platform,
  CAST(NULL AS TIMESTAMP) AS pay_create_time
FROM
  (
    SELECT
      *,
      PROCTIME() AS proctime
    from
      paimon_minio.order_dw.orders
  ) o
  LEFT JOIN paimon_minio.order_dw.product_catalog FOR SYSTEM_TIME AS OF o.proctime AS dim ON o.product_id = dim.product_id
UNION ALL
SELECT
  order_id,
  CAST(NULL AS STRING) AS user_id,
  CAST(NULL AS BIGINT) AS shop_id,
  CAST(NULL AS BIGINT) AS product_id,
  CAST(NULL AS STRING) AS order_product_catalog_name,
  CAST(NULL AS BIGINT) AS order_fee,
  CAST(NULL AS TIMESTAMP) AS order_create_time,
  CAST(NULL AS TIMESTAMP) AS order_update_time,
  CAST(NULL AS INT) AS order_state,
  pay_id,
  pay_platform,
  create_time
FROM
  paimon_minio.order_dw.orders_pay;

---

DWS

CREATE CATALOG paimon_minio WITH (
  'type' = 'paimon',
  'warehouse' = 's3://warehouse/wh',
  's3.endpoint' = 'http://192.168.1.243:9000',
  's3.access-key' = 'minio',
  's3.secret-key' = 'minio@123',
  's3.region' = 'us-east-1',
  's3.path.style.access' = 'true'
);
USE CATALOG paimon_minio;

SET 'execution.checkpointing.max-concurrent-checkpoints' = '3';
SET 'table.exec.sink.upsert-materialize' = 'NONE';
SET 'execution.checkpointing.interval' = '10s';
SET 'execution.checkpointing.min-pause' = '10s';

-- 为了同时计算用户视角的聚合表以及商户视角的聚合表，另外创建一个以用户 + 商户为主键的中间表。
CREATE TABLE IF NOT EXISTS paimon_minio.order_dw.dwm_users_shops (
    user_id STRING,
    shop_id BIGINT,
    ds STRING,
    payed_buy_fee_sum BIGINT COMMENT '当日用户在商户完成支付的总金额',
    pv BIGINT COMMENT '当日用户在商户购买的次数',
    PRIMARY KEY (user_id, shop_id, ds) NOT ENFORCED
) WITH (
    'merge-engine' = 'aggregation', -- 使用预聚合数据合并机制产生聚合表
    'fields.payed_buy_fee_sum.aggregate-function' = 'sum', -- 对 payed_buy_fee_sum 的数据求和产生聚合结果
    'fields.pv.aggregate-function' = 'sum', -- 对 pv 的数据求和产生聚合结果
    'changelog-producer' = 'lookup', -- 使用lookup增量数据产生机制以低延时产出变更数据
    -- dwm层的中间表一般不直接提供上层应用查询，因此可以针对写入性能进行优化。
    'file.format' = 'avro', -- 使用avro行存格式的写入性能更加高效。
    'metadata.stats-mode' = 'none' -- 放弃统计信息会增加OLAP查询代价（对持续的流处理无影响），但会让写入性能更加高效。
);

INSERT INTO paimon_minio.order_dw.dwm_users_shops
SELECT
    order_user_id,
    order_shop_id,
    DATE_FORMAT (pay_create_time, 'yyyyMMdd') as ds,
    order_fee,
    1 -- 一条输入记录代表一次消费
FROM paimon_minio.order_dw.dwd_orders
WHERE pay_id IS NOT NULL AND order_fee IS NOT NULL;

---

创建 dws_users 和 dws_shops

-- Step 2: ?? Paimon Catalog
CREATE CATALOG paimon_minio WITH (
  'type' = 'paimon',
  'warehouse' = 's3://warehouse/wh',
  's3.endpoint' = 'http://192.168.1.243:9000',
  's3.access-key' = 'minio',
  's3.secret-key' = 'minio@123',
  's3.region' = 'us-east-1',
  's3.path.style.access' = 'true'
);
USE CATALOG paimon_minio;
-- 用户维度聚合指标表。
CREATE TABLE IF NOT EXISTS paimon_minio.order_dw.dws_users (
  user_id STRING,
  ds STRING,
  payed_buy_fee_sum BIGINT COMMENT '当日完成支付的总金额',
  PRIMARY KEY (user_id, ds) NOT ENFORCED
) WITH (
  'merge-engine' = 'aggregation',
  -- 使用预聚合数据合并机制产生聚合表
  'fields.payed_buy_fee_sum.aggregate-function' = 'sum' -- 对 payed_buy_fee_sum 的数据求和产生聚合结果
  -- 由于dws_users表不再被下游流式消费，因此无需指定增量数据产生机制
);
-- 商户维度聚合指标表。
CREATE TABLE IF NOT EXISTS paimon_minio.order_dw.dws_shops (
  shop_id BIGINT,
  ds STRING,
  payed_buy_fee_sum BIGINT COMMENT '当日完成支付总金额',
  uv BIGINT COMMENT '当日不同购买用户总人数',
  pv BIGINT COMMENT '当日购买用户总人次',
  PRIMARY KEY (shop_id, ds) NOT ENFORCED
) WITH (
  'merge-engine' = 'aggregation',
  -- 使用预聚合数据合并机制产生聚合表
  'fields.payed_buy_fee_sum.aggregate-function' = 'sum',
  -- 对 payed_buy_fee_sum 的数据求和产生聚合结果
  'fields.uv.aggregate-function' = 'sum',
  -- 对 uv 的数据求和产生聚合结果
  'fields.pv.aggregate-function' = 'sum' -- 对 pv 的数据求和产生聚合结果
  -- 由于dws_shops表不再被下游流式消费，因此无需指定增量数据产生机制
);
SET
  'execution.checkpointing.max-concurrent-checkpoints' = '3';
SET
  'table.exec.sink.upsert-materialize' = 'NONE';
SET
  'execution.checkpointing.interval' = '10s';
SET
  'execution.checkpointing.min-pause' = '10s';
INSERT INTO
  paimon_minio.order_dw.dws_users
SELECT
  user_id,
  ds,
  payed_buy_fee_sum
FROM
  paimon_minio.order_dw.dwm_users_shops;
-- 以商户为主键，部分热门商户的数据量可能远高于其他商户。
-- 因此使用local merge在写入Paimon之前先在内存中进行预聚合，缓解数据倾斜问题。
INSERT INTO
  paimon_minio.order_dw.dws_shops
  /*+ OPTIONS('local-merge-buffer-size' = '64mb') */
SELECT
  shop_id,
  ds,
  payed_buy_fee_sum,
  1,
  -- 一条输入记录代表一名用户在该商户的所有消费
  pv
FROM
  paimon_minio.order_dw.dwm_users_shops;

ADS层 构建物化视图

starrocks 切换会 默认的catalog；

构建基于PAIMON的异步物化视图

查询结果：

相关语句如下：

CREATE MATERIALIZED VIEW ads_users_by_date
REFRESH ASYNC EVERY(INTERVAL 2 MINUTE)
PROPERTIES ("replication_num" = "1")
AS
select * from paimon_minio_catalog.order_dw.dws_users;
 

-------------------------------------

-- 查看 MV 状态
SHOW MATERIALIZED VIEWS LIKE 'ads_users_by_date';
 
-- 手动刷新（可选）
REFRESH MATERIALIZED VIEW ads_users_by_date;
 
-- 查询结果
SELECT * FROM ads_users_by_date;

1、同步任务 为 在命令行中提交的flink任务

2、分层建设任务 为在streampark上提供的3个任务

---

验证新数据的录入情况

捕捉业务数据库的变化

前面已完成了流式湖仓的构建，下面将测试流式湖仓捕捉业务数据库变化的能力。

PAIMON 中已有数据：

同步任务已启动。

在mysql中添加数据：

INSERT INTO orders VALUES
(100008, 'user_001', 12345, 3, 3000, '2023-02-15 17:40:56', '2023-02-15 18:42:56', 1),
(100009, 'user_002', 12348, 4, 1000, '2023-02-15 18:40:56', '2023-02-15 19:42:56', 1),
(100010, 'user_003', 12348, 2, 2000, '2023-02-15 19:40:56', '2023-02-15 20:42:56', 1);



INSERT INTO orders_pay VALUES
(2008, 100008, 1, '2023-02-15 18:40:56'),
(2009, 100009, 1, '2023-02-15 19:40:56'),
(2010, 100010, 0, '2023-02-15 20:40:56');

产看PAIMON中的数据

查看 dwd_orders, 数据已同步：

查看 dwm_users_shops, 数据已有变化：

返回 查看starrocks 内部表

原有数据：

数据已更新为：