Data modeling with datasets

Model your data in Sigma using a dataset.

A Sigma dataset is a collection of data based on a table in the CDW, and can include aggregations and other data manipulations. A dataset functions as a centralized, shareable data definition. Admins define a dataset that includes most or all of the data needed for analysis. Workbook creators use the dataset as the consistent source for visualizations, tables, and pivot tables.

Workbooks built on datasets inherit all changes to the underlying dataset. The benefit of creating workbooks using a dataset is that it is a single data source, ensuring that metrics are always calculated the same way. If a change to the calculations are needed, you update the dataset, and all workbooks built on it inherit the change.

Organizations have different use cases for datasets regarding access to data. These include:

  • Require all Creators to use datasets for workbooks and never directly access the CDW.
  • Primarily use the CDW for workbooks. Use a dataset to try out temporary changes, and move the changes to the CDW.
  • Use both, choosing either the CDW or a dataset depending on specific need.

Modeling data

Best practices

Annotate tables

Create datasets

Configure columns

Add links

The dataset worksheet

Join data

Sharing and permissions

Permission types

Share a dataset

Modify or revoke access

Respond to a request for access to content

Advanced modeling

Examine dataset queries

Dataset lineage

Edit warehouse data from Sigma

Dataset warehouse views

Dataset materialization

Modeling Best Practices Next