Data Connect Datasets
A Data Connect dataset selects a primary database table or configured form model, joins other sources, and organizes the resulting fields for charts and API clients.
Create A Dataset
- Open Data Connect → Datasets and create a dataset.
- Enter a name and select the data source.
- Select the primary table or form model from the saved source catalog.
- Configure joins if needed and create the dataset.
- Open the dataset, review its fields, and use Data preview to check the result.
A form model must have a usable data reading rule. You do not need to create form models for ordinary physical-table datasets.
Edit Sources And Relationships
Use Edit next to the refresh button to open the relationship editor.
Select the primary source, add joined sources, and choose the left source for each join. Join conditions match corresponding fields. The selector distinguishes dataset columns from form fields and can expose the source tables used by a form's reading rule, including available added fields.
| Join type | Result |
|---|---|
| Left | Keeps all left-side rows; missing right-side fields are empty |
| Inner | Keeps matching rows only |
| Right | Keeps right-side rows; missing left-side fields are empty |
| Full | Keeps rows from both sides, including unmatched rows |
MySQL and MariaDB do not support full joins in this configuration. A dataset supports up to 12 ordered joins. One-to-many joins repeat parent values; check totals and counts in the preview.
Applying the editor updates the draft. Save the dataset to make the change available to subsequent queries.
Manage Fields
Edit names, dimensions/measures, aggregation, visibility, and supported field options in the field list. A field's display name can change without changing its underlying identifier.
JSON/XML fields show their type. Configured scalar subfields are available to datasets and can be selected hierarchically in chart field pickers. Arrays are not expanded into chart subfields by this feature.
Add Computed Fields
Open Edit, then use the custom-column button in the relationship editor.
- Enter a stable field code and display name.
- Choose a supported data type.
- Choose an expression or Studio script value source.
- Configure the field and apply the editor.
- Save the dataset.
Expressions compute values from dataset fields using the supported formula language. They are not arbitrary SQL.
A Studio script field is a placeholder filled by the visualization's binding script. Selecting this source does not make API queries run a per-record script; the database query returns null for that placeholder, and an applicable chart script supplies its value.
Up to 80 computed fields can be configured. Keep the same field code when editing the field's type or value source if its API authorization should be preserved.
Refresh Structure And Resolve Invalid Fields
After source metadata changes:
- Refresh the source structure or definition scan as applicable.
- Open the dataset and click Refresh structure.
- Review added fields, invalid fields, and missing or invalid joins.
- Repair mappings, hide or remove obsolete fields as supported, and save.
- Preview the saved result.
Newly opening a dataset does not require saving if neither its editable configuration nor its reconciled structure has changed. Data preview requires the current draft and structure changes to be saved.
Use The Dataset
Select the saved dataset in Visualization. Chart previews can render field display values, including resolved reference labels, while filtering and matching continue to use the underlying values.
Configure internal member field and row permissions through the dataset permission settings. API roles are managed separately under the API service.
When granting a dataset visibility in the role permission editor, its enabled, grantable fields receive view permission in the draft. Review field permissions before saving. JSON/XML fields are grouped under their parent document in the field dialog; field visibility and row scope still apply.
Deleting a dataset used by a chart is blocked until that chart dependency is resolved. When a dataset is deleted, its bindings are also removed from API roles; other dataset bindings remain.