Module timely::dataflow::operators::aggregation
source · Expand description
Aggregation operators of various flavors
Two traits, Aggregate
and StateMachine
, which support the accumulation of streamed information.
Aggregate
accumulates records within times, and releases the accumulations once the time is complete.
StateMachine
responds to a sequence of keyed events, maintaining and updating a state for each key.
The user logic may produce output records for each transition, and optionally de-register the state to
clean up when appropriate.
The two methods are often combined, using first Aggregate
to reduce the volume of information, and then
StateMachine
to track an accumulation across timestamps.
Re-exports§
pub use self::aggregate::Aggregate;
pub use self::state_machine::StateMachine;
Modules§
- General purpose intra-timestamp aggregation
- General purpose state transition operator.