A user must implement Slowly Changing Dimensions (SCD Type 2) to retain historical changes.
Which action will achieve this?
SCD Type 2 maintains history by expiring the existing current dimension row and inserting a new version with its own effective dates. Snowflake streams provide row-level change data from a source or staging table, while tasks can process those changes with a MERGE; together, that logic can close the current row and create the replacement version.
MERGE
Community Discussion