Skip to main content
Each CSV file must contain a header row to specify which fields are included in the file. Each row thereafter will contain an action to be performed — such as creating a user — along with the fields associated with that object. Each type of object will have required fields and optional fields. Required fields are outlined below. An example header row and first action follows:

​
Duplicate Records

A batch should not contain multiple instances of the same record, such as two upsert events for the same account to update the balance. Individual records in a batch are processed asynchronously, and the order of execution is not guaranteed. Having duplicate records in the same batch will leave the record in an unpredictable state. Batches are processed sequentially. Ideally, you should ensure your previous batch has finished processing before initiating a new batch that will update records contained in the previous batch. However, it is permissible to upload the new batch after the previous batch has started since we order batch processing by PREFIX.*.YYYY-MM-DD.