What is the best practice to validate foreign keys on big timeseries Snowflake table?

Attempt to use “excpect_column_values_to_be in_set” utilizes 100 CPU for a very long time. Snowflake query to identify orphan records runs for secons only.