Functions that group data based on time
There are overlapping and non-overlapping windows
Hopping and Sliding are overlapping windows
Hopping Window
Hopes forwards in time using a fixed period
We need to specify the window size and the hop size. If both these values are same it is same as tumbling window
Overlapping of events/ windows can occur when using hopping window
Sliding Window
Does not aggregate values after a fixed time period
A new output is created when a new event is recorded or when event falls out of the window
An event can belong to more than one sliding window as well
Tumbling Window
Splits data into distinct time segments in which we perform aggregation operations
It is contagious, non-overlapping timing window and an event cannot belong to more than one window
It is the most commonly used window function
Session Window
Aggregates events that arrive at the same time
When no events occur no window is created
The SessionWindow() function takes three parameters the time unit, timeout size and the max duration size
Snapshot Window
Groups events that occur at the same timestamp
No window function is required
