How to Import CSV Data Into NetSuite Reliably

How to run a clean CSV import

1. 1Go to Setup > Import/Export > Import CSV Records and choose the record type (or Custom Record, or Transaction type such as Journal Entry).
2. 2Select the import behavior: Add (create new only), Update (existing only, matched by a key field), or Add or Update (upsert).
3. 3For Update or upsert, choose the match field carefully; Internal ID is the most reliable match, external ID or a unique custom field are next best, and Name is the riskiest because names are rarely unique.
4. 4In the field mapping step, map every required field explicitly; do not rely on NetSuite's auto-match for fields with similar but not identical column headers.
5. 5For list/record fields (customer, item, subsidiary, location), reference the exact value as it appears in NetSuite, or better, the internal ID in a dedicated column to avoid ambiguous text matches.
6. 6For transactions with line items (journal entries, sales orders), use the multi-line CSV format where each line item is a separate row sharing the same transaction-level key column, and check Multiple Line Items on the import.
7. 7Run the import first against a small test batch (5-10 rows) and check Setup > Import/Export > View CSV Import Status for row-level error detail before running the full file.
8. 8Save the mapping as an Import Map for repeat imports of the same file structure, under the mapping step's Save option.

Multi-line transaction imports

External ID,Date,Account,Debit,Credit,Memo
JE-1001,1/15/2026,Cash,1000.00,,Jan reclass
JE-1001,1/15/2026,AR,,1000.00,Jan reclass

Diagnosing failed rows

Common pitfalls

• !Matching Add or Update imports on Name instead of internal ID or external ID, risking updates to the wrong record when names collide.
• !Leaving list/reference field values as free text that does not exactly match NetSuite's stored display value, causing silent row failures.
• !Not checking Multiple Line Items on multi-row transaction imports, which causes each CSV row to become a separate single-line transaction instead of one multi-line transaction.
• !Importing without a small test batch first, then having to clean up hundreds of bad records after a full-file run.
• !Assuming CSV import respects the same field-level validation and workflow triggers as manual entry; some User Event script logic and workflow actions behave differently or are skipped depending on script deployment context settings.
• !Ignoring the import log's row-level errors and only checking the summary count, missing partial failures in a large file.