A fact table joins frequently with a dimension table on customer_id. To minimize data shuffling between nodes during joins, which distribution strategy fits?
A.EVEN distribution on both
EVEN spreads evenly — joins require shuffling.
B.KEY distribution on customer_id for both tables
Same distkey → co-located joins. Correct.
C.ALL distribution on the fact table
ALL duplicates the whole table per node — right for SMALL dims, wrong for a big fact table.
D.AUTO — let Redshift decide
AUTO can work but 'explicitly ensure co-location' is KEY.
Why this answer
KEY distribution co-locates rows with the same distkey value on the same node — joining both tables on customer_id becomes node-local (no shuffle). ALL copies small dimension tables to every node; EVEN round-robins (forces shuffles); AUTO lets Redshift optimize but the explicit join-locality answer is KEY.
Correct answer: B · Last reviewed 2026-11-05
Official AWS references
Practice this interactively
Answer without seeing the solution first in the Data Stores & Lakes practice session.
Practice Data Stores & Lakes