Sync ES

背景

canal 1.1.1版本之后, 内置增加客户端数据同步功能, Client适配器整体介绍: ClientAdapter

canal adapter 的 Elastic Search 版本支持6.x.x以上, 如需其它版本的es可替换依赖重新编译client-adapter.elasticsearch模块

ElasticSearch适配器

1 修改启动器配置: application.yml

canal.conf:
  canalServerHost: 127.0.0.1:11111
  batchSize: 500
  syncBatchSize: 1000
  retries: 0
  timeout:
  mode: tcp 
  srcDataSources:
    defaultDS:
      url: jdbc:mysql://127.0.0.1:3306/mytest?useUnicode=true
      username: root
      password: 121212
  canalAdapters:
  - instance: example 
    groups:
    - groupId: g1
      outerAdapters:
      - name: es
        hosts: 127.0.0.1:9300               # es 集群地址, 逗号分隔
        properties:
          mode: transport # or rest         # 可指定transport模式或者rest模式
          # security.auth: test:123456      # only used for rest mode
          cluster.name: elasticsearch       # es cluster name

adapter将会自动加载 conf/es 下的所有.yml结尾的配置文件

2 适配器表映射文件

修改 conf/es/mytest_user.yml文件:

dataSourceKey: defaultDS        # 源数据源的key, 对应上面配置的srcDataSources中的值
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doc名称
  _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 a
        left join role b on b.id=a.role_id
        left join (select user_id, group_concat(label order by id desc separator ';') as labels from label
        group 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:

1. 主表不能为子查询语句
2. 只能使用left outer join即最左表一定要是主表
3. 关联从表如果是子查询不能有多张表
4. 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)
5. 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
6. 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.

2.1.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"
                    }
                }
            }
        }
    }
}

2.2.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"
                    }
                }
            }
        }
    }
}

2.3.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"
                    }
                }
            }
        }
    }
}

2.4.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 label
        group 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"
                    }
                }
            }
        }
    }
}

2.5.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 label
        group by user_id) c on c.user_id=a.id

配置中使用:

objFields:
  labels: array:;

• 父子文档索引

es/customer.yml

......
esMapping:
  _index: customer
  _type: _doc
  _id: id
  relations:
    customer_order:
      name: customer
  sql: "select t.id, t.name, t.email from customer t"

es/order.yml

esMapping:
  _index: customer
  _type: _doc
  _id: _id
  relations:
    customer_order:
      name: order
      parent: customer_id
  sql: "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_time
        from 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"
          }
        }
      }
    }
  }
}

3 启动ES数据同步

启动canal-adapter启动器

bin/startup.sh

验证

1. 新增mysql mytest.user表的数据, 将会自动同步到es的mytest_user索引下面, 并会打出DML的log
2. 修改mysql mytest.role表的role_name, 将会自动同步es的mytest_user索引中的role_name数据
3. 新增或者修改mysql mytest.label表的label, 将会自动同步es的mytest_user索引中的labels数据