You are here:
Recipe Output to CSV
Use the Data Prep output node in CRM Analytics to save your prepared data as a data and schema file in a Salesforce BPO (storage database). You can then download the data locally using the public API as a CSV file. Use the CSV file to transfer prepared, cleaned, and improved data into your system for analysis, storage, or to inform business processes.
Push Data to the Salesforce Database
- Build your recipe with Data Prep and add an output node.
- Write the output node to CSV.
- Enter the user who downloads the file.
- (Optional) To output more information in the header, select Use fully
qualified names in CSV header.
The fully qualified name has more information about the source object to help you understand the data lineage, usually in the format [ObjectName].[APIName]. For example, the CSV header uses the fully qualified name Account.BillingCity instead of the column label BillingCity.
- (Optional) To set the maximum file size and maximum number of rows per partitioned section, select Specify CSV partition settings.
- Click Apply and save the recipe.
Pull Data from the Salesforce Database
To retrieve the data from the Salesforce database, use a tool to find the dataset ID. Download the file using the Salesforce CLI. Work with your Salesforce developer about using these tools and consult this Salesforce blog for a similar process.
CSV Output Considerations
Keep these behaviors in mind when working with the CSV output from the Data Prep recipe output node.
- An output connection is available only for recipes built with Data Prep.
- Stored data is removed after 48 hours.
- If the data exceeds the maximum file size or number of rows specified in the output node’s partition settings, more than one CSV file is created. All partitioned CSV files can function independently, and you can download and merge the files to recreate the dataset.
- The maximum size of an export job is 40 GB.
- The maximum export data size per day (sum of all exports) is 50 GB.
Exporting a Dataset
When you export a dataset, a DatasetExportEvent is generated. You can subscribe to receive DatasetExportEvents. See Subscribing to Platform Events. For event reference, see DatasetExportEvent.

