Skip to main content

Data Modeling

Tigris is quite flexible in terms of the different types of data models that are supported.

A collection's schema defines the structure of the documents that are stored in the collection. This schema maps to an entity or object in your application code. This is why you declare the data model as part of your application code with Tigris.

Tigris also enables you to easily modify your data models, all from your application code.

Embedded data model

Emebedded documents store related data together in a single document structure. Tigris allows you to have rich schemas that enable embedding related data in a single document using Object and Array field types.

Embedded models allow applications to retrieve and manipulate related data in a single database operation. The following examples demonstrate a collection that leverages this pattern by storing related data in a single data model:

Embedded data model

Relational data model

Sometimes there is a need to split related data across multiple collections. This is known as a normalized data model - this model stores relationships between data as references in the documents. Applications then resolve these references when accessing the data.

Below is an example that demonstrates a normalized data model:

Normalized data model

Data consistency and atomicity of operations

When using a relational data model, you would have data split across multiple collections. As a result, data modifications will likely result in modifications to multiple documents. As Tigris provides ACID transactions, you can safely modify multiple documents in a transactional manner with guaranteed atomicity.

info

With Tigris, there is no performance penalty for multi-document transactions. Hence, Tigris supports Embedded and relational data models equally well and leaves the choice to the user to decide what is best for their application use case.

Further reading

When choosing the right data model to use, always think about how your application works with the data and how the relationships between data is represented.

Checkout the language-specific reference sections to learn more about defining data models in your favorite language.