Aggregate over window(基于窗口的聚合(Aggregate over window))¶
Supported in: Streaming
Performs the specified aggregations on the data within a window, emitting outputs as specified by the provided trigger.
Transform categories: Aggregate
Declared arguments¶
- Aggregate expressions: List of aggregate expressions to evaluate over each window.
List\> - Dataset: Dataset to perform aggregations on.
Table - Window: Window defining how elements should be grouped.
Window - optional Accumulation mode: The accumulation mode for the window. Determines whether the window accumulates panes when the trigger fires or discards them.
Enum\ - optional Key by columns: Columns on which to partition the input by key. Each aggregation will be computed separately for each distinct key value.
Set\> - optional Trigger: Trigger defining when aggregation should be performed.
Trigger
中文翻译¶
基于窗口的聚合(Aggregate over window)¶
支持流式处理(Streaming)
对指定窗口内的数据执行聚合操作,并根据设定的触发条件(Trigger)输出结果。
转换类别:聚合(Aggregate)
声明参数¶
- 聚合表达式(Aggregate expressions): 需要在每个窗口内计算的聚合表达式列表。
List\> - 数据集(Dataset): 执行聚合操作的数据集。
Table - 窗口(Window): 定义数据分组方式的窗口。
Window - 可选 累积模式(Accumulation mode): 窗口的累积模式。决定触发条件触发时窗口是累积窗格(Pane)还是丢弃窗格。
Enum\ - 可选 分组列(Key by columns): 用于按键值对输入数据进行分区的列。每个不同的键值将分别计算聚合结果。
Set\> - 可选 触发条件(Trigger): 定义何时执行聚合操作的触发条件。
Trigger