A large fact table in Redshift is always joined to a customer dimension table on customer_id. Which distribution minimizes data movement across nodes during those joins?
A.DISTSTYLE EVEN on both tables
EVEN round-robins rows, so matching join keys land on different nodes and data must be redistributed.
B.DISTKEY(customer_id) on both tables
Correct. Matching customer_id rows co-locate, making joins collocated.
C.DISTSTYLE AUTO on the fact table only
AUTO picks a style over time but isn't deterministic for a known hot join — explicit DISTKEY is the exam answer.
D.DISTSTYLE ALL on the fact table
ALL copies the whole table to every node — fine for a small dimension, wasteful for a large fact table.
Why this answer
DISTKEY on the join column co-locates matching rows on the same slices, enabling collocated joins with no redistribution. A small dimension can also use DISTSTYLE ALL — but ALL on a large fact table would bloat every node.
Correct answer: B · Last reviewed 2026-10-26
Official AWS references
Practice this interactively
Answer without seeing the solution first in the Data Stores & Lakes practice session.
Practice Data Stores & Lakes