Question 1

Do I need an External ID field in Salesforce?

Accepted Answer

No, which surprises most people. Identity can be any Salesforce field that is filterable and otherwise eligible, so an account number, domain or email column already on the object can serve as the key without creating a dedicated External ID field first. Exact equality is standard, and string identities additionally support substring, domain and hostname matching, plus matching against a list of values.

Question 2

How are lookups and relationships set?

Accepted Answer

Either by writing the reference identifier directly, or through generated find by fields that resolve the related record for you. So a contact can be pointed at its account using the account's own identifier or an External ID style lookup, and Salesforce record identifiers do not have to be carried through the lakehouse. Where a reference does not resolve, you choose whether to fail the row, omit just the relationship, or skip the record.

Question 3

How should the model be shaped first?

Accepted Answer

Flat, with one row per record you intend to write. Salesforce fields hold individual values, so STRUCT, ARRAY and MAP columns should be resolved in the model SQL before they reach the mapping, otherwise a nested value has nowhere sensible to land. Doing that in SQL also keeps the mapping legible and means a change to what you write is a query change rather than a remapping exercise.

Question 4

What happens on a large first load?

Accepted Answer

The connector picks between the REST and Bulk APIs automatically based on volume, so a large backfill goes through Bulk rather than consuming per record calls, and operations pause at 80 percent of your daily API allowance. Two settings matter at scale: only updating when values differ costs extra reads, and orgs prone to row locking usually want a smaller Bulk batch size with serial processing.

Question 5

How do I connect Databricks to Salesforce?

Accepted Answer

Connect Databricks with a SQL warehouse hostname and a token, or an OAuth service principal, and connect Salesforce by authorizing against your org. The connected user's permissions decide which objects and fields the sync can write to, so authorize with someone who has the access you need. Then model the data, pick the target object, choose the field to match on, map your columns and set a schedule.

Connect Databricks with Salesforce

Connect Databricks and Salesforce with Polytomic

Match on any filterable field

API quota protected

Differential after backfill

Trusted by businesses worldwide

Step-by-step integration guide

Connect Databricks in Polytomic

Connect Salesforce in Polytomic

Create a sync between Databricks and Salesforce

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect Databricks with Salesforce?