实时数据湖：Flink CDC流式写入Hudi

点击上方蓝色字体，选择“设为星标”

回复"面试"获取更多惊喜

1. 环境准备

•Flink 1.12.2_2.11•Hudi 0.9.0-SNAPSHOT(master分支)•Spark 2.4.5、Hadoop 3.1.3、Hive 3.1.2

2. Flink CDC写入Hudi

MySQL建表语句如下

create table users
(
    id bigint auto_increment primary key,
    name varchar(20) null,
    birthday timestamp default CURRENT_TIMESTAMP not null,
    ts timestamp default CURRENT_TIMESTAMP not null
);
 
// 随意插入几条数据
insert into users (name) values ('hello');
insert into users (name) values ('world');
insert into users (name) values ('iceberg');
insert into users (id,name) values (4,'spark');
insert into users (name) values ('hudi');
 
select * from users;
update users set name = 'hello spark'  where id = 5;
delete from users where id = 5;

启动sql-client

$FLINK_HOME/bin/sql-client.sh embedded 
 
//1.创建 mysql-cdc
CREATE TABLE mysql_users (
                             id BIGINT PRIMARY KEY NOT ENFORCED ,
                             name STRING,
                             birthday TIMESTAMP(3),
                             ts TIMESTAMP(3)
) WITH (
      'connector' = 'mysql-cdc',
      'hostname' = 'localhost',
      'port' = '3306',
      'username' = 'root',
      'password' = '123456',
      'server-time-zone' = 'Asia/Shanghai',
      'database-name' = 'mydb',
      'table-name' = 'users'
      );
 
// 2.创建hudi表
CREATE TABLE hudi_users2
(
    id BIGINT PRIMARY KEY NOT ENFORCED,
    name STRING,
    birthday TIMESTAMP(3),
    ts TIMESTAMP(3),
    `partition` VARCHAR(20)
) PARTITIONED BY (`partition`) WITH (
    'connector' = 'hudi',
    'table.type' = 'MERGE_ON_READ',
    'path' = 'hdfs://localhost:9000/hudi/hudi_users2',
    'read.streaming.enabled' = 'true',
    'read.streaming.check-interval' = '1' 
);
 
//3.mysql-cdc 写入hudi ，会提交有一个flink任务
INSERT INTO hudi_users2 SELECT *, DATE_FORMAT(birthday, 'yyyyMMdd') FROM mysql_users;
 

Flink任务提交成功后可以查看任务界面

同时可以查看HDFS里的Hudi数据路径，这里需要等Flink 5次checkpoint（默认配置可修改）之后才能查看到这些目录，一开始只有.hoodie一个文件夹

在MySQL执行insert、update、delete等操作，当进行compaction生成parquet文件后就可以用hive/spark-sql/presto(本文只做了hive和spark-sql的测试)进行查询，这里需要注意下：如果没有生成parquet文件，我们建的parquet表是查询不出数据的。

3. Hive查询Hudi表

cd $HIVE_HOME
mkdir auxlib

然后将hudi-hadoop-mr-bundle-0.9.0-SNAPSHOT.jar拷贝过来

使用beeline登录hive

beeline -u jdbc:hive2://localhost:10000 -n hadoop hadoop

创建外部表关联Hudi路径，有两种建表方式

方式一：INPUTFORMAT是org.apache.hudi.hadoop.HoodieParquetInputFormat
这种方式只会查询出来parquet数据文件中的内容，但是刚刚更新或者删除的数据不能查出来
// 创建外部表
CREATE EXTERNAL TABLE `hudi_users_2`(               
   `_hoodie_commit_time` string,                    
   `_hoodie_commit_seqno` string,                   
   `_hoodie_record_key` string,                     
   `_hoodie_partition_path` string,                 
   `_hoodie_file_name` string,                      
   `id` bigint,                                     
   `name` string,                                   
   `birthday` bigint,                               
   `ts` bigint)                                     
 PARTITIONED BY (                                   
   `partition` string)                              
 ROW FORMAT SERDE                                   
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'  
 STORED AS INPUTFORMAT                              
   'org.apache.hudi.hadoop.HoodieParquetInputFormat' 
 OUTPUTFORMAT                                       
   'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat' 
 LOCATION                                           
   'hdfs://localhost:9000/hudi/hudi_users2'; 
 
方式二：INPUTFORMAT是org.apache.hudi.hadoop.realtime.HoodieParquetRealtimeInputFormat   
// 这种方式是能够实时读出来写入的数据，也就是Merge On Write，会将基于Parquet的基础列式文件、和基于行的Avro日志文件合并在一起呈现给用户。
 
CREATE EXTERNAL TABLE `hudi_users_2_mor`(               
   `_hoodie_commit_time` string,                    
   `_hoodie_commit_seqno` string,                   
   `_hoodie_record_key` string,                     
   `_hoodie_partition_path` string,                 
   `_hoodie_file_name` string,                      
   `id` bigint,                                     
   `name` string,                                   
   `birthday` bigint,                               
   `ts` bigint)                                     
 PARTITIONED BY (                                   
   `partition` string)                              
 ROW FORMAT SERDE                                   
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'  
 STORED AS INPUTFORMAT                              
   'org.apache.hudi.hadoop.realtime.HoodieParquetRealtimeInputFormat' 
 OUTPUTFORMAT                                       
   'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat' 
 LOCATION                                           
   'hdfs://localhost:9000/hudi/hudi_users2'; 
 
 
   
// 添加分区
alter table hudi_users_2 add if not exists partition(`partition`='20210414') location 'hdfs://localhost:9000/hudi/hudi_users2/20210414';
 
alter table hudi_users_2_mor add if not exists partition(`partition`='20210414') location 'hdfs://localhost:9000/hudi/hudi_users2/20210414';
   
// 查询分区的数据
select * from hudi_users_2 where `partition`=20210414;
select * from hudi_users_2_mor where `partition`=20210414;

INPUTFORMAT是org.apache.hudi.hadoop.realtime.HoodieParquetRealtimeInputFormat格式的表在hive3.1.2里面是不能够执行统计操作的

执行select count(1) from hudi_users3_mor where partition='20210414';

查看hive日志 tail -fn 100 hiveserver2.log

需要进行如下设置：set hive.input.format = org.apache.hudi.hadoop.hive.HoodieCombineHiveInputFormat ;具体原因参照这个issue：https://github.com/apache/hudi/issues/2813，或者阿里云技术文档：https://help.aliyun.com/document_detail/193310.html?utm_content=g_1000230851&spm=5176.20966629.toubu.3.f2991ddcpxxvD1#title-ves-82n-odd

再执行一遍依旧报错

但是在本地用hive-2.3.8执行成功了，社群里面的同学测试1.1版本的也报同样的错误，目前猜测是hive版本兼容性有关

4. Spark-SQL查询Hudi表

将hudi-spark-bundle_2.11-0.9.0-SNAPSHOT.jar拷贝到$SPAKR_HOME/jars，每个节点都拷贝一份

将hudi-hadoop-mr-bundle-0.9.0-SNAPSHOT.jar拷贝到$HADOOP_HOME/share/hadoop/hdfs下，每个节点都拷贝一份，然后重启hadoop

创建表，同样有两种方式

CREATE EXTERNAL TABLE `hudi_users3_spark`(               
   `_hoodie_commit_time` string,                    
   `_hoodie_commit_seqno` string,                   
   `_hoodie_record_key` string,                     
   `_hoodie_partition_path` string,                 
   `_hoodie_file_name` string,                      
   `id` bigint,                                     
   `name` string,                                   
   `birthday` bigint,                               
   `ts` bigint)                                     
 PARTITIONED BY (                                   
   `partition` string)                              
 ROW FORMAT SERDE                                   
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'  
 STORED AS INPUTFORMAT                              
   'org.apache.hudi.hadoop.HoodieParquetInputFormat' 
 OUTPUTFORMAT                                       
   'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat' 
 LOCATION                                           
   'hdfs://localhost:9000/hudi/hudi_users3';  
   
   alter table hudi_users3_spark add if not exists partition(`partition`='20210414') location 'hdfs://localhost:9000/hudi/hudi_users3/20210414';
  
  select * from hudi_users3_spark where `partition`='20210414';
 
// 创建可以实时读表数据的格式
CREATE EXTERNAL TABLE `hudi_users3_spark_mor`(               
   `_hoodie_commit_time` string,                    
   `_hoodie_commit_seqno` string,                   
   `_hoodie_record_key` string,                     
   `_hoodie_partition_path` string,                 
   `_hoodie_file_name` string,                      
   `id` bigint,                                     
   `name` string,                                   
   `birthday` bigint,                               
   `ts` bigint)                                     
 PARTITIONED BY (                                   
   `partition` string)                              
 ROW FORMAT SERDE                                   
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'  
 STORED AS INPUTFORMAT                              
   'org.apache.hudi.hadoop.HoodieParquetInputFormat' 
 OUTPUTFORMAT                                       
   'org.apache.hudi.hadoop.realtime.HoodieParquetRealtimeInputFormat' 
 LOCATION                                           
   'hdfs://localhost:9000/hudi/hudi_users3';  
  
  
  alter table hudi_users3_spark_mor add if not exists partition(`partition`='20210414') location 'hdfs://localhost:9000/hudi/hudi_users3/20210414';
  
  select * from hudi_users3_spark_mor where `partition`='20210414';

如果Spark-SQL读取实时Hudi数据，必须进行如下设置set spark.sql.hive.convertMetastoreParquet=false;

这里需要注意如果创建表的时候字段类型不对会报错，比如

CREATE EXTERNAL TABLE `hudi_users3_spark_mor`(               
   `_hoodie_commit_time` string,                    
   `_hoodie_commit_seqno` string,                   
   `_hoodie_record_key` string,                     
   `_hoodie_partition_path` string,                 
   `_hoodie_file_name` string,                      
   `id` string,                                     
   `name` string,                                   
   `birthday` string,                               
   `ts` string)                                     
 PARTITIONED BY (                                   
   `partition` string)                              
 ROW FORMAT SERDE                                   
   'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe'  
 STORED AS INPUTFORMAT                              
   'org.apache.hudi.hadoop.HoodieParquetInputFormat' 
 OUTPUTFORMAT                                       
   'org.apache.hudi.hadoop.realtime.HoodieParquetRealtimeInputFormat' 
 LOCATION                                           
   'hdfs://localhost:9000/hudi/hudi_users3';  

id 、ts、birthday都设置为String，会报下面错误。Spark-SQL想读取Hudi数据，字段类型需要严格匹配

5. 后续

目前使用小规模数据测试Flink CDC写入Hudi，后面我们准备用生产数据来走一波，看看Flink-CDC写入Hudi的性能和稳定性。

关于Presto避坑的小小指南

大数据方向另一个十年开启 |《硬刚系列》第一版完结

我写过的关于成长/面试/职场进阶的文章

我们在学习Spark的时候，到底在学习什么？

【面试&个人成长】2021年过半，社招和校招的经验之谈

Spark SQL重点知识总结

你好，我是王知无，一个大数据领域的硬核原创作者。

做过后端架构、数据中间件、数据平台&架构、算法工程化。

专注大数据领域实时动态&技术提升&个人成长&职场进阶，欢迎关注。