【问题标题】:Pyflink Table API Streaming Group WindowPyflink 表 API 流组窗口
【发布时间】:2021-06-17 21:31:39
【问题描述】:

我正在尝试对PyFlink 中的窗口进行一些聚合。但是我尝试它时得到A group window expects a time attribute for grouping in a stream environment. 错误。我在窗口定义和选择中都有一个时间属性。

我做错了什么?

from pyflink.common import Row
from pyflink.table import EnvironmentSettings, TableEnvironment, DataTypes
from pyflink.table.expressions import col 
from pyflink.table.window import Tumble

from datetime import datetime, date, time

    
# create a blink streaming TableEnvironment
env_settings = EnvironmentSettings.new_instance().in_streaming_mode().use_blink_planner().build()
table_env = TableEnvironment.create(env_settings)

schema = DataTypes.ROW([
    DataTypes.FIELD("rowtime", DataTypes.TIMESTAMP(3)), 
    DataTypes.FIELD("b", DataTypes.STRING()), 
    DataTypes.FIELD("c", DataTypes.STRING()), 
    ])
    

if __name__ == "__main__":

    t = table_env.from_elements([(datetime(1970, 1, 1, 0, 0, 0), 'Hi', 'Hello'),
                                 (datetime(1970, 1, 1, 1, 0, 0), 'Hi', 'hi'),
                                 (datetime(1970, 1, 1, 2, 0, 0), 'Hi2', 'hi'),
                                 (datetime(1970, 1, 1, 3, 0, 0), 'Hi', 'Hello'),
                                 (datetime(1970, 1, 1, 4, 0, 0), 'Hi', 'Hello')], schema=schema)

    print(
            t.window(Tumble.over("1.hour").on("rowtime").alias("w")) \
              .group_by("w")\
              .select(col("w").start, col("w").end, col("w").rowtime)
    )

Traceback (most recent call last):
  File "test2.py", line 31, in <module>
    .select(col("w").start, col("w").end, col("w").rowtime)
  File "/home/user/venv/lib/python3.6/site-packages/pyflink/table/table.py", line 1290, in select
    return Table(self._j_table.select(to_expression_jarray(fields)), self._t_env)
  File "/home/user/venv/lib/python3.6/site-packages/py4j/java_gateway.py", line 1286, in __call__
    answer, self.gateway_client, self.target_id, self.name)
  File "/home/user/venv/lib/python3.6/site-packages/pyflink/util/exceptions.py", line 147, in deco
    return f(*a, **kw)
  File "/home/user/venv/lib/python3.6/site-packages/py4j/protocol.py", line 328, in get_return_value
    format(target_id, ".", name), value)
py4j.protocol.Py4JJavaError: An error occurred while calling o46.select.
: org.apache.flink.table.api.ValidationException: A group window expects a time attribute for grouping in a stream environment.
    at org.apache.flink.table.operations.utils.AggregateOperationFactory.validateStreamTimeAttribute(AggregateOperationFactory.java:329)
    at org.apache.flink.table.operations.utils.AggregateOperationFactory.validateTimeAttributeType(AggregateOperationFactory.java:309)
    at org.apache.flink.table.operations.utils.AggregateOperationFactory.getValidatedTimeAttribute(AggregateOperationFactory.java:302)
    at org.apache.flink.table.operations.utils.AggregateOperationFactory.createResolvedWindow(AggregateOperationFactory.java:267)
    at org.apache.flink.table.operations.utils.OperationTreeBuilder.windowAggregate(OperationTreeBuilder.java:257)
    at org.apache.flink.table.api.internal.TableImpl$WindowGroupedTableImpl.select(TableImpl.java:783)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at org.apache.flink.api.python.shaded.py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
    at org.apache.flink.api.python.shaded.py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
    at org.apache.flink.api.python.shaded.py4j.Gateway.invoke(Gateway.java:282)
    at org.apache.flink.api.python.shaded.py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
    at org.apache.flink.api.python.shaded.py4j.commands.CallCommand.execute(CallCommand.java:79)
    at org.apache.flink.api.python.shaded.py4j.GatewayConnection.run(GatewayConnection.java:238)
    at java.lang.Thread.run(Thread.java:748

【问题讨论】:

    标签: python apache-flink pyflink


    【解决方案1】:

    我认为时间戳格式有问题或尝试使用TIMESTAMP_LTZ 列(接受纪元时间,即1618989564564)。

    从文档中,它说:

    Flink 支持在 TIMESTAMP 列上定义事件时间属性和 TIMESTAMP_LTZ 列。如果源中的时间戳数据是 表示为年-月-日-时-分-秒,通常为字符串 没有时区信息的值,例如2020-04-15 20:13:40.564, 建议将事件时间属性定义为 TIMESTAMP 专栏”

    【讨论】:

      【解决方案2】:

      我不是专家,但我根据您的代码做了一个工作示例。

      我必须使用数据流 api 并格式化日期时间并定义水印。

      #!/usr/bin/env python
      # -*- coding: utf-8 -*-
      
      from datetime import datetime
      
      from pyflink.common import Types
      from pyflink.datastream import StreamExecutionEnvironment
      from pyflink.table import (
          DataTypes,
          Schema,
          StreamTableEnvironment,
      )
      from pyflink.table.expressions import col
      from pyflink.table.window import Tumble
      
      env = StreamExecutionEnvironment.get_execution_environment()
      t_env = StreamTableEnvironment.create(env)
      
      
      def main():
          ds = env.from_collection(
              collection=[
                  (datetime(1970, 1, 1, 0, 0, 0).isoformat().replace('T', ' '), 'Hi', 'Hello'),
                  (datetime(1970, 1, 1, 1, 0, 0).isoformat().replace('T', ' '), 'Hi', 'hi'),
                  (datetime(1970, 1, 1, 2, 0, 0).isoformat().replace('T', ' '), 'Hi2', 'hi'),
                  (datetime(1970, 1, 1, 3, 0, 0).isoformat().replace('T', ' '), 'Hi', 'Hello'),
                  (datetime(1970, 1, 1, 4, 0, 0).isoformat().replace('T', ' '), 'Hi', 'Hello'),
              ],
              type_info=Types.ROW([Types.STRING(), Types.STRING(), Types.STRING()]),
          )
          schema = Schema.new_builder() \
              .column_by_expression("rowtime", "CAST(f0 AS TIMESTAMP(3))") \
              .column("f1", DataTypes.STRING()) \
              .column("f2", DataTypes.STRING()) \
              .watermark("rowtime", "rowtime - INTERVAL '3' SECOND") \
              .build()
          the_table = t_env.from_data_stream(ds, schema)
          res = the_table.to_pandas()
          print(res)
      
          res = the_table \
              .window(Tumble.over("1.hour").on("rowtime").alias("w")) \
              .group_by("w") \
              .select(
                  col("w").start.alias('start_time'),
                  col("w").end.alias('end_time'),
                  col("w").proctime.alias('proctime'),
              ) \
              .to_pandas()
          print(res)
      
      
      if __name__ == '__main__':
          main()
      

      【讨论】:

        猜你喜欢
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 1970-01-01
        • 2020-09-13
        • 1970-01-01
        相关资源
        最近更新 更多