How to map CSV columns before importing into Airtable linked records

CSV imports into Airtable often fail for reasons that are not obvious from the import screen. The file may look clean, the headers may be readable, and the preview may even show the right rows. The real problem is usually the mapping between source columns, Airtable field types, and linked record behavior.

This is the workflow I use before importing operational data into Airtable, especially when the base has linked records, select fields, IDs, or messy exports from another system.

Start with a mapping table

Before touching the importer, map the CSV in a small table:

CSV column	Airtable field	Field type	Required?	Cleanup rule
Contact ID	Source ID	Single line text	Yes	Preserve exactly
Company	Company	Linked record	Yes	Match primary field
Status	Status	Single select	No	Normalize labels
Notes	Notes	Long text	No	Keep raw text
Created At	Created date	Date	No	Convert to one date format

This table makes the import decision explicit. It also helps you avoid changing the Airtable schema while you are already inside the upload flow.

Separate identifiers from names

A common mistake is treating every readable value as a display label. In practice, imports usually contain two different kinds of columns:

stable identifiers, such as customer IDs, ticket IDs, order IDs, or source system IDs
human-readable labels, such as company names, contact names, status labels, or category names

Preserve stable IDs exactly. Do not trim leading zeros, convert them to numbers, or merge them into display fields. If something goes wrong later, those IDs are what let you compare the imported Airtable records against the original export.

Be careful with linked records

Linked records deserve special attention. If the CSV value does not match the linked table's primary field, the import can look correct but still fail to create the relationship you expected.

For example, if your CSV has a Company column with values like "Acme Inc" but the linked Companies table uses "Acme" as the primary field, Airtable may not match the existing company record. You can end up with blank links, duplicate linked records, or records that look valid but are not connected correctly.

Before importing linked records, check:

whether the linked table already contains the target records
whether the CSV value exactly matches the linked table primary field
whether you want Airtable to create new linked records automatically
whether duplicate names exist in the linked table
whether the import should use a text staging field first

For risky imports, I usually import into a temporary text field first, review the values, then convert or map them into linked records after the mismatch cases are visible.

Normalize select fields before import

Single select and multi-select fields can create messy option lists if the CSV contains small variations.

These values may all represent the same status:

Active
active
ACTIVE
Active customer
Current

Decide the final Airtable labels before import. Otherwise, you may create duplicate select options and have to clean the base after the data is already live.

Test with five representative rows

Do not test with the first five rows unless they represent the full data shape. Pick rows that include:

one normal record
one record with a linked record value
one record with missing optional fields
one record with date or number fields
one record with edge-case text, such as commas, line breaks, or unusual characters

This small test catches most mapping errors before the full import creates hundreds or thousands of records.

Create a post-import review view

After import, create a review view that filters for likely problems:

blank required fields
unmatched linked records
duplicate source IDs
unexpected select values
invalid dates
records created during the test import

This view is useful because import errors are not always visible in the upload screen. Some issues only become obvious after the records exist in the base.