A DataRaptor Extract pulls Salesforce data into a JSON shape. The Designer walks you through the source object, the relationship traversal, and the output mapping; the result is a callable mapper that any OmniStudio element can invoke.
- Open the DataRaptor Designer
OmniStudio app, then DataRaptors tab (now labeled Data Mappers in Summer '24 orgs and later), then New. Pick Extract as the interface type.
- Name and configure the source
Give the DataRaptor a clear name (Account_GetWithContacts is better than Test1). On the Extract tab, pick the primary object and add any SOQL filters in the Extract Filter Definition.
- Define the output structure
On the Output tab, map source fields to output paths. Nest output paths with dot notation (Account.Owner.Name) and create lists with array notation (Account.Contacts[]). The Designer's preview pane shows the shape as you build.
- Add formulas and conditions
Use the Formula column for derived fields and the Conditions column to include or exclude rows. Functions like CONCAT, IF, and DATEVALUE behave like Salesforce formula fields with minor argument differences.
- Set caching if appropriate
On the Options tab, enable Cache Response and pick a TTL. Use this for reference data and read-heavy patterns; skip it for anything user-personal or rapidly changing.
- Save and test
Click Save, then use the Preview tab to invoke the DataRaptor with sample input. The response panel shows the shaped output and timing. Once it looks right, the DataRaptor is callable from any Integration Procedure or OmniScript.
Reads from Salesforce objects with full transformation. The default for read patterns that need joining or shaping.
Pure transformation on input JSON, no database access. Used to convert between payload formats.
Writes a JSON payload to one or more objects, with upsert by External Id supported.
Fast read of one object at a time, field-level mapping only, no formulas. The right choice for high-volume reads with simple shapes.
- DataRaptor Turbo Extract does not support formulas, conditions, or relationship traversal. If you need any of those, drop back to a regular Extract.
- Response caching is keyed on the input parameters. A DataRaptor with no input parameters caches one response for all callers, which is rarely what you want.
- DataRaptor Load runs as a single DML batch. Hitting the 10,000-record DML limit fails the whole load, not just the offending rows. Split very large loads across multiple invocations.
- The Summer '24 rename to Data Mapper does not affect existing references. You can keep using DataRaptor in conversation and in custom code without breaking anything.