Some integrations, like Airtable, require an identifier unique to each record / row.
Every data set uses a different identifier field. Many use the standardized, convenient and plain "id", but many others use custom identifier fields like "AccountID" or "TransactionID" etc.
So you must lookup this information manually by (1) Asking ChatGPT what the unique identifier field is for your data set from the integration you are using, or (2) Looking at API documentation to find it.
Example:
Ask ChatGPT "What is the unique identifier field is for Plaid Transactions?"
Alternatively, look at Plaid's API Documentation:
https://plaid.com/docs/api/products/transactions/#transactionsget
In this example "transaction_id" the identifier.
Question: Why doesn't Flatly do this for me?
Answer: There is no way for Flatly to predetermine which identifier applies to each data set. (many integrations X many datasets). Furthermore, if Flatly applies the wrong identifier, the flattened data could be erroneous or malformed.