In Snowflake, a role is an access control entity that can be created as part of a hierarchy within an account. Roles are used to grant and manage privileges in a structured and scalable manner.
Understanding Roles:
Roles are logical entities that group privileges together.
They are used to control access to securable objects like tables, views, warehouses, and more.
Role Hierarchy:
Roles can be organized into a hierarchy, allowing for the inheritance of privileges.
A role higher in the hierarchy (parent role) can grant its privileges to a lower role (child role), simplifying privilege management.
Creating Roles:
Roles can be created using the CREATE ROLE command.
You can define parent-child relationships by granting one role to another.
Example Usage:
CREATE ROLE role1;
CREATE ROLE role2;
GRANT ROLE role1 TO role2;
In this example, role2 inherits the privileges of role1.
Benefits:
Simplifies privilege management: Hierarchies allow for efficient privilege assignment and inheritance.
Enhances security: Roles provide a clear structure for managing access control, ensuring that privileges are granted appropriately.
Reference:
Snowflake Documentation: Access Control in Snowflake
Snowflake Documentation: Creating and Managing Roles

In Snowflake, the CONNECT BY clause allows users to perform recursive queries. Recursive queries are used to process hierarchical or tree-structured data, such as organizational charts or file systems. The CONNECT BY clause is used in conjunction with the START WITH clause to specify the starting point of the hierarchy and the relationship between parent and child rows.
Reference:
Snowflake Documentation: Hierarchical Queries

Type predicates in Snowflake are used to determine if a value in a VARIANT column is of a particular data type. This is useful when working with semi-structured data stored in VARIANT columns, as it allows for data type validation and conditional processing based on the data type.
Reference:
Snowflake Documentation: Type Predicates

When cloning a database in Snowflake, the clone includes all privileges on the database as well as existing child objects within the database, such as schemas, tables, views, etc. However, it does not include future child objects or privileges on schemas within the database2.
Reference = [COF-C02] SnowPro Core Certification Exam Study Guide, Snowflake Documentation

The PARTITION BY <expression> parameter option in the COPY INTO <location> command is used to split the output into multiple files based on the distinct values of the specified expression. This feature is particularly useful for organizing large datasets into smaller, more manageable files and can help with optimizing downstream processing or consumption of the data. For example, if you are unloading a large dataset of transactions and use PARTITION BY DATE(transactions.transaction_date), Snowflake generates a separate output file for each unique transaction date, facilitating easier data management and access.
This approach to data unloading can significantly improve efficiency when dealing with large volumes of data by enabling parallel processing and simplifying data retrieval based on specific criteria or dimensions.
Reference:
Snowflake Documentation on Unloading Data: COPY INTO <location>