How to insert null values using salesforce Data Loader?

By default, Data Loader treats blank cells as “don’t touch this field”. If you put an empty value in your CSV and run an Update, the existing field value stays unchanged — your “delete this value” intent silently fails.

To actually null out fields you have to flip one switch.

The setting

1. Open Data Loader.
2. Settings → Settings.
3. Tick Insert Null Values.
4. Click OK.

From now on, every blank cell in your CSV is interpreted as “set this field to null”.

Bulk API equivalent

If you use the Bulk API option in Data Loader (or any other Bulk API client), the equivalent toggle is the #N/A literal in the CSV cell. The Bulk API treats #N/A as an explicit null request:

Id,Phone
0035g00000ABCDE,#N/A

The phone number on that record gets cleared.

Why this matters

The default behaviour exists for a reason: most loads are partial — you want to update some fields and leave others alone. Blank-equals-null would be a footgun if it were the default.

But for a cleanup pass — “every record where IndustrySector is invalid, blank the field” — you need the setting on. Forgetting it is a classic Data Loader frustration: you run the job, success.csv shows 50,000 rows updated, and not one of those fields actually changed.

Things you still can’t null this way

• Required fields. If a field is Required at the schema level, attempting to null it returns REQUIRED_FIELD_MISSING.
• Required fields on layout but optional in schema — these can be nulled via the API; the layout requirement only enforces in the UI.
• System fields (Id, CreatedDate, SystemModstamp, etc.) — you can’t write these, null or otherwise.
• Master-detail lookups. Master-detail fields require a value by definition; nulling raises REQUIRED_FIELD_MISSING.
• Auto-Number fields — read-only after creation.

A safer workflow for nulling

Before running a bulk null operation in production:

1. Export the records first with all fields you’re about to clear. This is your backup.
2. Tick Insert Null Values in Settings.
3. Test in a sandbox with a small subset (50 rows).
4. Verify the result in the UI on a few records.
5. Run in production with the success/error files captured and checked in.
6. Untick Insert Null Values after the job if other admins share the install — leaving it on can mangle their next routine load.

A trap

If you’re using Upsert and want to create new records with some fields null, the Insert Null Values setting matters there too. Without it, an Upsert insert leaves the field blank only if Salesforce’s default is blank — your CSV’s empty cell is ignored.

Verified against: Data Loader Guide — Configuring Data Loader, Bulk API Developer Guide, Metadata API Developer Guide. Last reviewed 2026-05-17 for Spring ‘26 release.