The simple option is creating the cluster without specifying any device. If that method is chosen, then Regatta will map all tables and indexes for you.

To create the described table, enter the following SQL command:

 CREATE TABLE employees (
 employee_key INT PRIMARY KEY INDEX,
 employee_name VARCHAR(40) NOT NULL,
 employee_salary INT,
 employee_department VARCHAR(50) NOT NULL)

Enter SQL command SHOW TABLES:

SHOW TABLES

In response, the table details will be displayed. We can see the employee table with all the columns and the device mappings. If we do not specify the mapping, the system will distribute the table rows across all the devices and choose a single random device for the indexes.

Notes:

  • When looking at the mapping, you can see that the same device appears three times. This is the out-of-the-box default mapping, which delivers high performance.
  • This mapping is not necessarily the most performant, but it is the simplest behavior, and it allows someone to easily get started (which is the purpose of this document).
  • When there is more than one device used for indexing, we need to define range separators that will define how to map the indexes to the different devices. It is up to the user to specify this mapping. In the default mapping the indexes are mapped to a single device and therefore there is no need for range separators, so it defaults to no range separators.
  • For further details, please refer to the Table creation with default data distribution** **section.