Design databases and tables for an ApsaraDB for SelectDB instance - ApsaraDB for SelectDB

When you store your business data in an ApsaraDB for SelectDB instance, you must design important table properties based on your business requirements. This helps you create a high-performance and easy-to-maintain table schema. The following table describes the important table properties of a SelectDB instance.

Table property	Required	Description	References
Data model	Yes	Different data models are suitable for different business scenarios. The Unique model supports uniqueness constraints on the primary key. This allows you to update data in a flexible and efficient manner. The Duplicate model uses the append data write mode and is suitable for high-performance analysis of detailed data. The Aggregate model supports data pre-aggregation and is suitable for data aggregation and statistics scenarios.	Data models
Bucket	Yes	Buckets are used to distribute data to different nodes in a cluster to manage and query large amounts of data by leveraging the capabilities of a distributed system.	Partitioning and bucketing
Partition	No	Partitions are used to divide a raw table into multiple child tables based on specified fields, such as time and region. This facilitates data management and query by partition and accelerate queries.	Partitioning and bucketing
Index	No	Indexes are used to quickly filter or locate data. This greatly improves the query performance.	Indexes

If this is the first time you design a table for a SelectDB instance, we recommend that you understand the key points in database and table design. After you have an overall understanding of the key points in table design, you can read the relevant details in depth.