Ttyjfyk

I am using the simple-salesforce Python package to upsert records into our Salesforce instance. I am finding that each bulk API call counts as 6 API requests. Is this correct? Is there a way to reduce this whilst performing this operation.

The function I am using to perform the upsert:

def bulk_upsert(self, object_name, data_set, identity_field):


 upsert_response = SFBulkType(object_name, self.bulk_con.connect.bulk_url,
 self.bulk_con.connect.headers, self.bulk_con.connect.session).upsert(data_set, identity_field)

 return upsert_response

The object name is the custom object created in Salesforce for our instance, the dataset is a list of dictionaries containing 200 records, the identity field is the external id for the Salesforce object.

I establish the Salesforce connection (sf) and then loop through requested_chunks which is a list of the data to upsert broken into 200 records each.

for request in request_chunks:

 upsert_response = sf.bulk_upsert(sf_object, request, sf_identity) 

I am finding that each time through this loop it is costing 6 credits. I have done a comparison with the integration recommended by simple-salesforce and that also costs 6 calls per bulk upsert. Any thoughts?

Is your python lib uses Bulk API or simple record insert/update?
If it uses Salesforce Bulk API it needs 1 request to create Bulk Job, 1 request to upload records, 1 request to mark job as posted.
And needs to poll request to track job status if records are processed

The Bulk API is designed for bulk uploads of data. If you're importing less than 1,000 records, you're definitely wasting API calls. Six calls sounds about right; it definitely requires at least three just to initialize, upload, and start the Bulk API call, plus more calls to wait for completion. Use the normal upsert API call if you want to upload smaller batches of data (e.g. 200 records at a time).