Strip Optional Null Fields from JSON Before Processing

A profile edit form sends all fields; only name and bio changed. Stripping null yields a body that won't wipe email or phone.

Input (update.json):
{
  "name": "Ada",
  "email": null,
  "phone": null,
  "bio": "Engineer"
}

Strip Values (targets: null · Deep on):
{
  "name": "Ada",
  "bio": "Engineer"
}

2 would-be overwrites avoided.

The user removed their avatar, which IS the change. Don't tick null so avatarUrl stays in the body as an explicit clear.

Input:
{ "avatarUrl": null, "theme": null, "name": "Ada" }

Strip Values (targets: empty strings only · null UNCHECKED):
{ "avatarUrl": null, "theme": null, "name": "Ada" }

Both nulls survive because null wasn't a target.
Send avatarUrl:null deliberately; handle theme separately.

Turning off a flag must reach the server. false is never stripped, so it stays in the PATCH body alongside the cleaned nulls.

Input:
{ "verified": false, "middleName": null }

Strip Values (targets: null · Deep on):
{ "verified": false }

The false update is preserved; the unset middleName is dropped.

Updating a nested settings object; Deep mode strips its optional nulls too so the nested PATCH is also minimal.

Input:
{
  "id": 5,
  "settings": { "locale": "en", "timezone": null }
}

Strip Values (targets: null · Deep on):
{
  "id": 5,
  "settings": { "locale": "en" }
}

If your stack is Prisma, convert null to undefined so the field is skipped rather than nulled — the code mirror of stripping.

// After stripping or in the client:
await prisma.user.update({
  where: { id },
  data: {
    name: updates.name,
    email: updates.email ?? undefined  // null -> skip
  }
});

Under JSON Merge Patch and many PATCH endpoints, an explicit null deletes/clears the field. If that's your intent, do NOT strip it — leave the null checkbox off, or send the clearing field in a separate request.

Setting a flag to false is a real change and is never stripped. The body keeps "verified": false so the server applies it. There's no falsy mode that could drop it.

"bio": "" (clear bio) or "tags": [] (remove all tags) are real updates. Leave empty strings / empty arrays unticked so they survive into the PATCH body.

Stripping produces an object without the key. Prisma's update wants undefined for 'skip', which JSON can't represent. Convert null to undefined in code (?? undefined) — stripping models the intent but the runtime fix is the JS side.

PUT replaces the whole resource, so a stripped field may be interpreted as 'removed' rather than 'unchanged'. Only strip for PATCH/merge-patch style partial updates, not full-replace PUTs.

A trailing comma or BOM throws on parse and the run fails. Fix the file with json-format-fixer before stripping.

With Deep unchecked, nested objects keep their nulls. For a nested partial update, turn Deep on so the nested resource is also minimised.

It's a no-op — JSON files can't contain undefined. Use null here and handle undefined in your application code.

Free caps at 2 MB, Pro at 100 MB. Update payloads are tiny, so this rarely applies — but a bulk array of updates could hit it.

If you want to drop or keep specific keys regardless of value (e.g. always exclude internalId), use json-key-filter instead of value-based stripping.