Supported Datasets and Models

Event tables

Mitzu currently supports two main types of event tables in the data warehouse.

  • Single-event tables - tables that contain a single event type. Typical examples of this are page_views, click_events tables. These tables have two mandatory columns user_id and event_time (can be named differently)

  • Multi-event tables - tables that contain multiple event types. The main difference between these tables and the single event tables is that these tables contain an event_type or event_name column in addition to the user_id and event_time. Typical examples for these tables are web_events, all_app_events, subscription_events, etc.

Modeling event properties

Single-event tables - typically, columns of these tables represent event properties.

Multi-event tables - most often, columns represent event properties. However, this leads to many missing column values if events don't support a certain event property. A more modern approach is to use a single column with a dynamic column type, such as json and map.

Mitzu supports all the above-listed options. If your data model is not support please contact for help.

User tables

User tables are tables that contain information about the users of the company. In these tables, every row represents a single user of the business. Each column in this table represents user properties.

Mitzu currently supports user tables.

Group tables

Group tables are similar to user tables as they contain information about the company to which the users belong. These tables are most common for B2B companies. Each row represents an account the users belong to.

Mitzu currently supports group tables.

Generic star schema support

We plan to support a generic star schema in Mitzu in the future. For now, we limited this feature to two dimensions.

Last updated