Question

A data analyst runs a dashboard query and inspects its Query Profile. The profile shows a single node labeled 'QUERY RESULT REUSE' and reports that the query executed in under a second with no bytes scanned. The analyst is surprised because the underlying table contains billions of rows. Which conclusion is correct about what happened?

Accepted Answer

Snowflake served the results entirely from the result cache because an identical query was run within the last 24 hours and the underlying data had not changed.

Answer

The warehouse local disk (SSD) cache held all the micro-partitions, so the compute nodes read data directly from local disk instead of remote storage.

Answer

The metadata cache in the cloud services layer computed the result from micro-partition min/max statistics without scanning any data.

Answer

The query benefited from clustering and search optimization, allowing Snowflake to prune all but a few micro-partitions before returning results.