Shuffle

Groups incoming pairs by key, emitting each key together with the list of values that shared it.

Use this when

• You are aggregating tagged sensor or telemetry readings by type.
• You need to batch values by a partition key before downstream reduction or analysis.
• You want a lightweight group-by without writing a custom aggregation component.

What it does

Consumes (K, V) pairs and groups them by key. Emits (K, [V]) for each distinct key. Inputs ("a", 1) ("b", 2) ("a", 3) ("b", 4) ("a", 5) produce ("a", [1, 3, 5]) ("b", [2, 4]).

Inputs

Type	Description
[(K, V)]	A list of keyed pairs.

Outputs

Type	Description
[(K, [V])]	One entry per distinct key, values grouped.

Parameters

This transformation takes no parameters.

Works best with

• Sensor network components that tag readings with a type or source ID.
• Downstream reducers, averagers, or per-group analytics.

Caveats

• Requires a finite input list — will not group across unbounded streams without upstream batching.
• Value order within each group follows input order but is not otherwise guaranteed across groups.