What's new in Sigma

Admin

Account type permissions for reports

The following account type permissions are now available to enable access to reports:

View reports: Enables users to view reports shared with them.
Create, edit, and publish reports: Enables users to create new reports, manage existing ones, and share published versions.

For more information, see Account type permission availability matrix.

Audit log support for Azure UK (UK South) region

Sigma now supports audit log events for organizations hosted in the Azure UK (UK South) region.

See Enable or disable audit logging for more information.

AI

New supported Databricks model family for inference

Sigma now uses the GTE v1.5 model family instead of the BGE v1.5 model family for inference when using Databricks as the warehouse model connection.

See all supported models in Configure a warehouse-hosted model as AI provider.

AI apps

Input table edit log updates

Sigma now creates edit logs (also known as write-ahead logs or WALs) at the individual input table level and stores each edit log in the same schema as its corresponding input table data.

For non-OAuth connections, edit logs continue to use the same user configurations or service account credentials to access and write to the destination schema. OAuth connections, however, no longer require a service account and instead use the same OAuth credentials to write input table data and the corresponding edit log.

For more information, see How input table data is handled and Write-back architecture for input tables with OAuth.

Reports

Create pixel-perfect reports (Beta)

Create reports in Sigma to generate predictable and pixel-perfect exports. Reports allow you to:

Match compliance and branding requirements with granular font, layering and positioning capabilities.
Organize your analyses into distinct numbered pages.
Add and paginate table and input table data up to 10,000 rows.
Send tailored reports to the right users with controls and email bursting. Reports can also be embedded. For more information about reports, see Reports overview (Beta) and Tutorial: Build and export a sales report.

Workbook elements

Language support for specific Python libraries (Beta)

The Python element in Sigma now supports inline error linting and autocomplete for the following libraries:

pandas, numpy, scikit-learn, requests, and scipy
Connection-specific libraries:
- For Databricks: pyspark library completions as well as spark and dbutils globals
- For Snowflake: snowflake-snowpark-python as well as session global

Language support is on by default, but can be turned off by each user. For more details, see Write and run Python code in Sigma.

Bug fixes

For chart interactions, data is now correctly scoped in the Selection action variable. By including values from all columns relevant to a chart selection and excluding values that are not applicable, this fix addresses formula evaluation errors and unexpected variable output.
The Update row(s) action in a version tagged workbook now references the correct workbook ID (that of the tagged version instead of the published version) to run as expected.
When migrating a dataset, the migration tool attempted to migrate deleted datasets that referenced the dataset.
When migrating a dataset joined to another dataset sourced from SQL, the SQL was not migrated to the new data model.