Creating from business objects

Select a primary business object (for example, Account or Invoice). Then add related objects using the relationship builder. The interface uses business terms rather than SQL syntax, making it easy to understand how your data connects.

By default, all records from your primary object are retained, even if related records are missing. This ensures you see complete results.

Creating from data query

For advanced use cases, you can build a dataset from a saved data query or write a new one. The query becomes the governed dataset that reports can use. This allows you to define custom data logic once, then reuse it across multiple reports without exposing the raw SQL.

Datasets and saved queries

A dataset is distinct from a saved data query, although datasets can be created based on saved queries. Here is how they relate:

• A saved data query is a stored SQL query that you can execute to retrieve raw data.

• A dataset created from a saved query functions as a governed copy. It takes the query logic and wraps it in dataset governance, including field definitions, metrics, and filters.

• If two datasets have identical metadata (same fields, metrics, and filters), they are considered equivalent datasets.

When you create a report directly from a saved query without first creating a dataset, the system automatically generates a dataset for that report using dual-write logic. This ensures governance is always in place. You can view all available saved queries when creating or modifying datasets.

If you create multiple datasets from the same saved query or with the same underlying metadata, you may accumulate unused datasets. You or your administrator can clean these up from the datasets UI if they are no longer needed.

The recommended workflow is to first create a dataset from an existing dataset (when applicable), then select Saved Queries to create new datasets with specific customizations. This approach keeps your dataset inventory organized and aligned with your reporting needs.

Creating from an existing dataset

You can create a derived dataset based on another dataset. This lets you extend or customize an existing definition with additional fields, metrics, or filters while keeping a clear lineage to the original.

Dataset filters

Dataset-level filters are global constraints that apply to every report using this Dataset. For example, a filter might exclude inactive accounts or archived transactions.