The correct answer is D. The existing micro-partitions will not be changed or impacted .
Important correction:
The provided answer B is not correct.
Snowflake stores table data in immutable micro-partitions. Because micro-partitions are immutable, Snowflake does not physically rewrite all existing micro-partitions immediately when a column is dropped. Instead, metadata is updated so that the dropped column is no longer visible or accessible in the table definition.
Why D is correct:
Dropping a column is a metadata operation. The underlying existing micro-partitions are not immediately rewritten to remove the column data. Snowflake’s immutable micro-partition design avoids expensive full-table rewrites for this type of operation.
Why the other options are incorrect:
A. Snowflake does not immediately create new micro-partitions containing only the remaining columns.
B. Existing micro-partitions are immutable and are not updated in place.
C. Snowflake does not automatically create a new table and rewrite all remaining data when a column is dropped.
Official Snowflake documentation reference:
Snowflake documentation describes micro-partitions as immutable storage units. Table metadata operations, such as dropping columns, do not immediately rewrite existing micro-partitions.
[Reference: Snowflake Documentation — Micro-partitions and data clustering; Snowflake Documentation — ALTER TABLE ... DROP COLUMN; SnowPro Core Study Guide — Snowflake Architecture., ========================]
