如何在卡夫卡中拥有对日期的现场类型支持

Question

使用debezium-mongodb-connector我设法将我的收藏推向kafka，我遇到的唯一问题是我的收藏中的 date在我的一种藏品中使用这种格式2019-05--14T23：25：34.703 00：00，并没有以相同的格式推到主题，而是我得到了类似的东西1560708085175。

这是我的Debezium Connector命令connect-standalone /etc/kafka/connect-standalone.properties /etc/kafka/connect-mongodb-source.properties 这是我的MongoDB集合示例。

{"_id":"5cdb4e6ed767ba70593e2aa8","sender":"5cdb43db4505956efc70ba03","receiver":"5cdb43db4505956efc70ba03","receiverWalletId":"5cdb43db4505956efc70ba04","status":"succes","type":"topup","amount":200000,"totalFee":0,"createdAt":"2019-05-14T23:25:34.703Z","updatedAt":"2019-05-14T23:25:35.132Z","__v":0,"details":"none."}

这是我的kafka主题示例。

{"schema":{"type":"struct","fields":[{"type":"string","optional":true,"field":"sender"},{"type":"string","optional":true,"field":"receiver"},{"type":"string","optional":true,"field":"receiverWalletId"},{"type":"string","optional":true,"field":"status"},{"type":"string","optional":true,"field":"type"},{"type":"int32","optional":true,"field":"amount"},{"type":"int32","optional":true,"field":"totalFee"},{"type":"int64","optional":true,"field":"createdAt"},{"type":"int64","optional":true,"field":"updatedAt"},{"type":"int32","optional":true,"field":"__v"},{"type":"string","optional":true,"field":"from"},{"type":"string","optional":true,"field":"orderId"},{"type":"string","optional":true,"field":"id"}],"optional":false,"name":"mongo_conn.digi.transactions"},"payload":{"sender":"5cef970ca2e9c273c655483","receiver":"5cef970ca2e9c27355c483","receiverWalletId":"5cef970ca2e9c27556c484","status":"pending","type":"topup","amount":6000,"totalFee":0,"createdAt":1560708024322,"updatedAt":1560708024753,"__v":0,"from":"smt","orderId":"d7a97581-9d18-79cd-8b09-16e400a43714","id":"5d0683b8be4af834abe3cf58"}}

这是我的连接mongodb-source.properties

name=mongodb-source-connector
connector.class=io.debezium.connector.mongodb.MongoDbConnector
mongodb.hosts=repracli/**.**.**.***27017
mongodb.name=mongo_conn
initial.sync.max.threads=1
tasks.max=1
transforms=unwrap
transforms.unwrap.type=io.debezium.connector.mongodb.transforms.UnwrapFromMongo$
transforms.unwrap.operation.header=true

Answer 1

对于多种转换，您将需要类似的东西：

transforms=unwrap,convert1,convert2
transforms.unwrap.type=io.debezium.connector.mongodb.transforms.UnwrapFromMongoDbEnvelope
transforms.unwrap.operation.header=true
transforms.convert1.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert1.target.type=string
transforms.convert1.field=createdAt
transforms.convert1.format=yyyy-MM-dd HH:mm:ss ZZZ
transforms.convert2.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert2.target.type=string
transforms.convert2.field= *updatedAt*
transforms.convert2.format=yyyy-MM-dd HH:mm:ss ZZZ

Answer 2

debezium在OPLOG中存储时以格式流式传输数据。自时代以来，日期看起来像毫秒的时间戳。

您可以编写SMT(https://cwiki.apache.org/confluence/display/kafka/kip-66: single message transage transforms for kafka kafka connect(您首选字符串表示的请求字段。

如果您研究了org.bson.BsonDateTime，您会发现它确实是long值。

Answer 3

已解决

name=mongodb-source-connector
connector.class=io.debezium.connector.mongodb.MongoDbConnector
mongodb.hosts=repracli/**.**.**.***:27017
mongodb.name=mongo_conn
initial.sync.max.threads=1
tasks.max=1
transforms=unwrap,convert,convert2,convert3,convert4
transforms.unwrap.type=io.debezium.connector.mongodb.transforms.UnwrapFromMongoDbEnvelope
transforms.unwrap.operation.header=true
transforms.convert.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert.target.type=string
transforms.convert.field=createdAt
transforms.convert.format=yyyy-MM-dd HH:mm:ss ZZZ
transforms.convert2.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert2.target.type=string
transforms.convert2.field=updatedAt
transforms.convert2.format=yyyy-MM-dd HH:mm:ss ZZZ
transforms.convert3.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert3.target.type=string
transforms.convert3.field=created_at
transforms.convert3.format=yyyy-MM-dd HH:mm:ss ZZZ
transforms.convert4.type=org.apache.kafka.connect.transforms.TimestampConverter$Value
transforms.convert4.target.type=string
transforms.convert4.field=updated_at
transforms.convert4.format=yyyy-MM-dd HH:mm:ss ZZZ