MapReduce Join
A MapReduce join implements joins in a distributed computing framework: Map tags the source, Shuffle groups by join key, Reduce merges.
Purpose
To understand the distributed join execution model and analyze shuffle costs of big-data joins.
Mechanism
Map: each row emits (join_key, (tag, row)). Shuffle: rows with the same key gather at one Reducer. Reduce: group by tag, then nested-loop or hash join. Reduce-side joins are general but shuffle-heavy; map-side joins require broadcasting the small table.
Use cases
- Big-data joins
- MapReduce optimization
- Shuffle cost analysis