Micro-partition
Snowflake splits every table into small micro-partitions and records each column’s value range, so queries can skip data — if the values are grouped well.
Snowflake splits every table into small micro-partitions and records each column’s value range, so queries can skip data — if the values are grouped well.
Clustering depth measures how much micro-partitions overlap on a column. Lower means more data can be skipped — but query performance is still the real test.