文章目录

数据同步ElasticSearch
- 单表基本配置
- 适配器映射文件详细介绍（单表、多表映射介绍）
- - 单表映射索引示例sql
  - 单表映射索引示例sql带函数或运算操作
  - 多表映射(一对一, 多对一)索引示例sql
  - 多表映射(一对多)索引示例sql
  - 其它类型的sql示例
  - 注意事项

本文详细介绍Canal 配置保存 ElasticSearch

Canal从零配置使用参考：https://blog.csdn.net/zhangshenghang/article/details/120361721

数据同步ElasticSearch

我们接着在之前配置Hbase基础上直接修改配置，实现同时同步ElasticSearch

单表基本配置

1.修改启动器配置 {canal-apapter}/conf/application.yml

server:port: 8081
logging:level:com.alibaba.otter.canal.client.adapter: DEBUGcom.alibaba.otter.canal.client.adapter.hbase: DEBUG
spring:jackson:date-format: yyyy-MM-dd HH:mm:sstime-zone: GMT+8default-property-inclusion: non_null
canal.conf:# tcp kafka rocketMQ rabbitMQ canal-server运行的模式，TCP模式就是直连客户端，不经过中间件。kafka和mq是消息队列的模式mode: tcp 
#  flatMessage: truezookeeperHosts: syncBatchSize: 1retries: 0timeout: 1000accessKey:secretKey:consumerProperties:# canal tcp consumer 指定canal-server的地址和端口canal.tcp.server.host: 127.0.0.1:11111canal.tcp.zookeeper.hosts: 127.0.0.1:2181canal.tcp.batch.size: 1canal.tcp.username:canal.tcp.password:srcDataSources: # 数据源配置，从哪里获取数据defaultDS: # 指定一个名字，在ES的配置中会用到，唯一url: jdbc:mysql://127.0.0.1:3306/test2?useUnicode=trueusername: rootpassword: *****canalAdapters:- instance: example # canal instance Name or mq topic name 指定在canal配置的实例名称groups:- groupId: g1 outerAdapters:- name: logger
#      - name: rdb
#        key: mysql1
#        properties:
#          jdbc.driverClassName: com.mysql.jdbc.Driver
#          jdbc.url: jdbc:mysql://127.0.0.1:3306/mytest2?useUnicode=true
#          jdbc.username: root
#          jdbc.password: 121212
#      - name: rdb
#        key: oracle1
#        properties:
#          jdbc.driverClassName: oracle.jdbc.OracleDriver
#          jdbc.url: jdbc:oracle:thin:@localhost:49161:XE
#          jdbc.username: mytest
#          jdbc.password: m121212
#      - name: rdb
#        key: postgres1
#        properties:
#          jdbc.driverClassName: org.postgresql.Driver
#          jdbc.url: jdbc:postgresql://localhost:5432/postgres
#          jdbc.username: postgres
#          jdbc.password: 121212
#          threads: 1
#          commitSize: 3000- name: hbaseproperties:hbase.zookeeper.quorum: sangfor.abdi.node3,sangfor.abdi.node2,sangfor.abdi.node1hbase.zookeeper.property.clientPort: 2181zookeeper.znode.parent: /hbase-unsecure- name: es7 # config目录下的子目录名称hosts: 192.168.168.2:9300 # 127.0.0.1:9200 for rest modeproperties:mode: transport # or rest
#          # security.auth: test:123456 #  only used for rest modecluster.name: my_application
#        - name: kudu
#          key: kudu
#          properties:
#            kudu.master.address: 127.0.0.1 # ',' split multi address

2.ElasticSearch 表映射文件

# 指定数据源，这个值和adapter的application.yml文件中配置的srcDataSources值对应。
dataSourceKey: defaultDS
# 指定canal-server中配置的某个实例的名字，不同实例对应不同业务
destination: example
# 组ID ，tcp方式这里填写空，不要填写值，不然可能会接收不到数据
groupId: 
# ES的mapping（映射）
esMapping:# ES索引名称_index: testsync2# ES标示文档的唯一标示，通常对应数据表中的主键ID字段_id: _id
#  upsert: true
#  pk: id
# 数据表每个字段映射到表中的具体名称，不能重复sql: "select a.id as _id, a.name,a.age,a.age_2,a.message,a.insert_time from testsync as a"
#  objFields:
#    _labels: array:;
#  etlCondition: "where a.c_time>={}"commitBatch: 10

3 重启服务

bin/restart.sh

写入数据

INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());
INSERT INTO testsync(id,name,age,insert_time) values(UUID(),UUID(),2,now());

查看adapter日志

2021-09-20 13:53:07.279 [pool-1-thread-1] INFO  c.a.o.canal.client.adapter.logger.LoggerAdapterExample - DML: {"data":[{"id":"05fabf89-19d7-11ec-bbe0-708cb6f5eaa6","name":"05fabfb4-19d7-11ec-bbe0-708cb6f5eaa6","age":2,"age_2":null,"message":null,"insert_time":1632117185000}],"database":"test2","destination":"example","es":1632117185000,"groupId":"g1","isDdl":false,"old":null,"pkNames":["id"],"sql":"","table":"testsync","ts":1632117187278,"type":"INSERT"}
2021-09-20 13:53:07.286 [pool-1-thread-1] DEBUG c.a.o.c.client.adapter.hbase.service.HbaseSyncService - DML: {"data":[{"id":"05fabf89-19d7-11ec-bbe0-708cb6f5eaa6","name":"05fabfb4-19d7-11ec-bbe0-708cb6f5eaa6","age":2,"age_2":null,"message":null,"insert_time":1632117185000}],"database":"test2","destination":"example","es":1632117185000,"groupId":"g1","isDdl":false,"old":null,"pkNames":["id"],"sql":"","table":"testsync","ts":1632117187278,"type":"INSERT"}
2021-09-20 13:53:07.287 [pool-1-thread-1] DEBUG c.a.o.canal.client.adapter.es.core.service.ESSyncService - DML: {"data":[{"id":"05fabf89-19d7-11ec-bbe0-708cb6f5eaa6","name":"05fabfb4-19d7-11ec-bbe0-708cb6f5eaa6","age":2,"age_2":null,"message":null,"insert_time":1632117185000}],"database":"test2","destination":"example","es":1632117185000,"groupId":"g1","isDdl":false,"old":null,"pkNames":["id"],"sql":"","table":"testsync","ts":1632117187278,"type":"INSERT"} 
Affected indexes: testsync2

查看ElasticSearch数据
在这里插入图片描述
至此写入ElasticSearch、Hbase成功

适配器映射文件详细介绍（单表、多表映射介绍）

${adapter}/conf/es7/xxx.yml

dataSourceKey: defaultDS        # 源数据源的key, 对应上面配置的srcDataSources中的值
outerAdapterKey: exampleKey     # 对应application.yml中es配置的key 
destination: example            # cannal的instance或者MQ的topic
groupId:                        # 对应MQ模式下的groupId, 只会同步对应groupId的数据
esMapping:_index: mytest_user           # es 的索引名称_type: _doc                   # es 的type名称, es7下无需配置此项_id: _id                      # es 的_id, 如果不配置该项必须配置下面的pk项_id则会由es自动分配
#  pk: id                       # 如果不需要_id, 则需要指定一个属性为主键属性# sql映射sql: "select a.id as _id, a.name as _name, a.role_id as _role_id, b.role_name as _role_name,a.c_time as _c_time, c.labels as _labels from user aleft join role b on b.id=a.role_idleft join (select user_id, group_concat(label order by id desc separator ';') as labels from labelgroup by user_id) c on c.user_id=a.id"
#  objFields:
#    _labels: array:;           # 数组或者对象属性, array:; 代表以;字段里面是以;分隔的
#    _obj: object               # json对象etlCondition: "where a.c_time>='{0}'"     # etl 的条件参数commitBatch: 3000                         # 提交批大小

sql映射说明:

sql支持多表关联自由组合, 但是有一定的限制:

主表不能为子查询语句
只能使用left outer join即最左表一定要是主表
关联从表如果是子查询不能有多张表
主sql中不能有where查询条件(从表子查询中可以有where条件但是不推荐, 可能会造成数据同步的不一致, 比如修改了where条件中的字段内容)
关联条件只允许主外键的’='操作不能出现其他常量判断比如: on a.role_id=b.id and b.statues=1
关联条件必须要有一个字段出现在主查询语句中比如: on a.role_id=b.id 其中的 a.role_id 或者 b.id 必须出现在主select语句中

Elastic Search的mapping 属性与sql的查询值将一一对应(不支持 select *), 比如: select a.id as _id, a.name, a.email as _email from user, 其中name将映射到es mapping的name field, _email将映射到mapping的_email field, 这里以别名(如果有别名)作为最终的映射字段. 这里的_id可以填写到配置文件的 _id: _id映射.

单表映射索引示例sql

select a.id as _id, a.name, a.role_id, a.c_time from user a

该sql对应的es mapping示例:

{"mytest_user": {"mappings": {"_doc": {"properties": {"name": {"type": "text"},"role_id": {"type": "long"},"c_time": {"type": "date"}}}}}
}

单表映射索引示例sql带函数或运算操作

select a.id as _id, concat(a.name,'_test') as name, a.role_id+10000 as role_id, a.c_time from user a

函数字段后必须跟上别名, 该sql对应的es mapping示例:

{"mytest_user": {"mappings": {"_doc": {"properties": {"name": {"type": "text"},"role_id": {"type": "long"},"c_time": {"type": "date"}}}}}
}

多表映射(一对一, 多对一)索引示例sql

select a.id as _id, a.name, a.role_id, b.role_name, a.c_time from user a 
left join role b on b.id = a.role_id

注:这里join操作只能是left outer join, 第一张表必须为主表!!
该sql对应的es mapping示例:

{"mytest_user": {"mappings": {"_doc": {"properties": {"name": {"type": "text"},"role_id": {"type": "long"},"role_name": {"type": "text"},"c_time": {"type": "date"}}}}}
}

多表映射(一对多)索引示例sql

select a.id as _id, a.name, a.role_id, c.labels, a.c_time from user a 
left join (select user_id, group_concat(label order by id desc separator ';') as labels from labelgroup by user_id) c on c.user_id=a.id

注:left join 后的子查询只允许一张表, 即子查询中不能再包含子查询或者关联!!

该sql对应的es mapping示例:

{"mytest_user": {"mappings": {"_doc": {"properties": {"name": {"type": "text"},"role_id": {"type": "long"},"c_time": {"type": "date"},"labels": {"type": "text"}}}}}
}

其它类型的sql示例

geo type

select ... concat(IFNULL(a.latitude, 0), ',', IFNULL(a.longitude, 0)) AS location, ...

复合主键

select concat(a.id,'_',b.type) as _id, ... from user a left join role b on b.id=a.role_id

数组字段

select a.id as _id, a.name, a.role_id, c.labels, a.c_time from user a 
left join (select user_id, group_concat(label order by id desc separator ';') as labels from labelgroup by user_id) c on c.user_id=a.id

配置中使用:

objFields:labels: array:;

对象字段

select a.id as _id, a.name, a.role_id, c.labels, a.c_time, a.description from user a

配置中使用:

objFields:description: object

其中a.description字段内容为json字符串

父子文档索引
es/customer.yml

......
esMapping:_index: customer_type: _doc_id: idrelations:customer_order:name: customersql: "select t.id, t.name, t.email from customer t"

es/order.yml

esMapping:_index: customer_type: _doc_id: _idrelations:customer_order:name: orderparent: customer_idsql: "select concat('oid_', t.id) as _id,t.customer_id,t.id as order_id,t.serial_code as order_serial,t.c_time as order_timefrom biz_order t"skips:- customer_id

mapping示例:

{"mappings":{"_doc":{"properties":{"id": {"type": "long"},"name": {"type": "text"},"email": {"type": "text"},"order_id": {"type": "long"},"order_serial": {"type": "text"},"order_time": {"type": "date"},"customer_order":{"type":"join","relations":{"customer":"order"}}}}}
}

注意事项

多表映射时，主表数据必须插入，如果只插入子表不插入主表，数据无法同步到ElasticSearch；相反只插入主表，子表不进行插入，数据是可以同步到ElasticSearch的
多表映射时，如果主表关联id写入后，子表再进行修改之前的关联的id为我们主表写入的id，数据是无法同步到ElasticSearch中的。