Datagraph Object

Datagraph Object Data Node with a published Object and bound parameter; the Result limit control is hidden by default

Overview

The Datagraph Object Data Node executes the Live version of a Datagraph Object and exposes the returned rows to downstream Nodes. It is the only way to consume the Datagraph model inside a Workflow. Unlike a Datasource Node, which reads a single Datasource, this Node runs the joined SQL prepared in the Object. Filtering, joining, and ordering therefore live in the Object, not in the Node.

The Datagraph is in beta, and the Node is restricted to tech admins. Other users do not see it in the Data menu of the editor.

Adding the Node

Hover the Data button in the editor toolbar and select Datagraph Object. See Data Nodes for the general procedure.

Data menu of the Workflow editor with the Datagraph Object option at the bottom of the list

Configuration

Field	Description
Object	Select an Object from Select a published object. Only Objects with a Live version appear; a sole Live Object is preselected.
Object parameters	One field per parameter in the selected Object’s Live definition. Each parameter’s short description appears above its input.

The empty selector message is No published object is available yet. Open object in Datahub opens the Object in a new tab and current account scope. Changing the selected Object clears the parameter values already set, because parameter names belong to the Object definition.

Publishing a new Object version that renames or adds a parameter does not update Nodes already configured. Re-open each Node to bind the new parameters, otherwise mandatory values are missing at execution time.

Only tech admins can configure Datagraph object query nodes.

The optional Result limit field is hidden by default. It appears only when a tech admin enables Datagraph Object: result limit in the Workflows settings → Features tab.

Binding Parameters

Each parameter is bound either to a static value or to a value produced upstream in the Workflow, through the Variable Picker. The input control follows the declared parameter type and shows its short description above the field.

Parameter type	Control
`string`, `id`	Text value or Workflow value
`number`	Numeric value or Workflow value
`datetime`, `datetime_iso`	Date value or Workflow value
`boolean`	Text value or Workflow value

Datagraph Object parameter input showing its type and short description, with available Workflow values

A parameter takes a single value. Multiple values for one parameter are not supported. Mandatory parameters must be bound. If a mandatory parameter has no value at execution time, the Object is rejected with an invalid parameter error rather than returning an empty result.

When the Object declares a client_id parameter and the Workflow resolves exactly one tracked person, that identifier is used for client_id if you left the field empty. Any value you bind explicitly always wins.

Suggested Parameter Values

An Object can declare a suggested value and a list of allowed values per parameter — see Suggested values. Only a suggested value pre-fills the matching input: allowed values never fill anything, they only turn the input into a select or multi-select. Pre-fill only applies when you select the Object in the drawer, and only to parameters that are still empty. Values you already set are never overwritten, and an existing Node’s parameters are not re-filled when you open it.

Suggestion source	Pre-filled value
From context	A link to the Context variable in the Context Node.
From Team attribute	The values resolved on the Workflow’s Teams: `string` takes the first value, `string[]` the whole list. A Team without the attribute leaves the parameter empty.
Fixed value	The declared static value.

Team attributes are inherited like Team scopes:

The closest Team wins per attribute.
A Team’s list replaces the inherited one — lists never merge.
The climb stops at a Team that excludes ancestor inheritance.
Two parents at the same level are taken in Team order.

An attribute set on a parent Team is not inherited by a child Team that excludes ancestor inheritance — set it directly on that Team for its Workflows to use it.

A pre-filled parameter shows a Suggested label. Clicking it clears the value; the label also disappears as soon as you edit the value.

Datagraph Object Node parameters pre-filled from suggestions, each marked with the Suggested label

A parameter whose Object declares allowed values gets a select (single value) or a multi-select (list) instead of the Variable Picker. Clearing that select — or clicking Suggested — switches the parameter back to the Variable Picker for the rest of the drawer session. A parameter already linked to a variable, or holding a value outside the list, keeps the Variable Picker.

A suggestion only writes a normal parameter value: it changes nothing at execution time.

Example: Team Attributes

Attributes are set per Team in the Team create/edit modal — see Team attributes. Take this Team tree:

Europe (parent) sets brand = [Acme, Acme Pro] and excluded_categories = [outlet, clearance].
France, a child of Europe, sets no attributes and inherits them.
Germany, a child of Europe, excludes ancestor inheritance and has no attributes.
Spain, a child of Europe, sets its own excluded_categories = [outlet] and inherits the rest.

The Object declares brand (string, suggestion From Team attribute brand) and excluded_categories (string[], suggestion From Team attribute excluded_categories). Pre-fill in a Workflow scoped to each Team:

Workflow’s Team	`brand`	`excluded_categories`
`France`	`Acme` (first value of the inherited list)	`outlet`, `clearance` (whole inherited list)
`Spain`	`Acme` (inherited)	`outlet` — its own list replaces the inherited one, with no merge
`Germany`	stays empty	stays empty — inheritance stops at `Germany`; set the attributes on `Germany` itself

Output

The Node returns a list of rows. Each row holds the output columns inferred by the Object, with text, numeric, or empty values. Those columns are exposed to downstream Nodes in the Variable Picker, under the name of the Object, and can feed a Content or a Condition. The canvas Node displays the name of the selected Object, so a Workflow using several Objects stays readable.

Datagraph Object Node on the Workflow canvas, showing the name of the selected Object

Worked Example

A property listings Datagraph joins users, listings, recommendations, and market prices. The Object listings_matching_a_user_search declares one parameter, user_id, and returns one row per matching listing with its reference, surface, city, and price. In the Workflow:

A URL Parameter Node captures user_id from the call made by the sending tool.
A Datagraph Object Node selects listings_matching_a_user_search and binds user_id to that URL Parameter. If enabled, set Result limit to 6.
The Content iterates over the returned rows to build a six-listing block.

Because the joins live in the Object, the same Node configuration works for every recipient, and the model can be corrected without touching the Workflow.

Operational Notes

If the Result limit field is enabled, keep it aligned with what the Content renders. Rows that are fetched but never displayed still cost execution time.
Test the Object with Run query in the Object editor before wiring it into a Workflow, using the same parameter values.
Relations with a low join match produce sparse results. Review them in the Relations tab on the Schema page before blaming the Node.
An Object whose Datasources are protected by row-level filters cannot be executed from a Workflow.