Snowflake - Build and Architect Data Pipelines Using AWS - Lab - Selecting Your Clustering Keys

The video tutorial discusses the considerations and best practices for clustering tables in a cloud data warehouse, specifically Snowflake. It emphasizes the importance of understanding analytical workloads and the ratio of queries to DML operations. The tutorial provides guidelines for selecting clustering keys, highlighting the significance of cardinality and the order of columns. Practical steps for clustering in Snowflake, including creating and altering tables, are also covered.

What is a key consideration when deciding to cluster a table?

• The table should be small in size.

• The table should have frequent DML operations.

• The table should be large, ideally multiple terabytes.

• The table should have low query selectivity.

Why is it important to understand the ratio of queries to DML operations?

• To minimize the need for clustering.

• To increase the number of DML operations.

• To ensure frequent updates.

• To determine the need for clustering based on query frequency.

What should be avoided when selecting clustering keys?

• Columns with text values.

• Columns with moderate cardinality.

• Columns with high or low cardinality.

• Columns with date values.

Which type of column is considered a good candidate for clustering keys?

• Timestamp columns

• Text columns

• Gender columns

• Date columns

What is the recommended order for specifying columns in clustering keys?

• From high to low cardinality

• From low to high cardinality

• Alphabetical order

• Random order

What is a safe approach when experimenting with clustering keys?

• Directly alter the original table.

• Use random columns for clustering.

• Avoid using any clustering keys.

• Create a clone of the table to test changes.

What happens to clustering keys when a table is cloned?

• They are removed.

• They are duplicated.

• They are retained.

• They are randomized.