What should be considered when you are asked to select the distribution key columns for a very large fact table?
A. As many columns as possible.
B. Columns that contain many nulls.
C. Columns contain few unique values.
D. Columns used to join to other large tables.
