# mark_processed

> Source: https://docs.clonepartner.com/reference/steps/mark-processed/

<!-- sources: src/steps/mark-processed-step.ts, src/steps/step-connector-utils.ts, src/steps/types.ts, references/envoy/step-types.md -->

# mark_processed

Updates the source row or document that produced the current item — typically to set a processed flag, the destination ID, or a content hash — and passes the item through unchanged.

## When to use it

Use `mark_processed` as the last write in a pipeline whose source is a database you can write to, so a later delta run can filter out completed records at the source. When the destination write itself needs dedup and change tracking, the `writeback` block of [sync_record](/reference/steps/sync-record/) covers the same update in one step.

## Configuration

| Key | Type | Required | Default | Description |
|---|---|---|---|---|
| `connector` | string | Yes | — | SQL (`sqlite`, `mssql`, `mysql`, `postgres`) or MongoDB Connector slot. Other types fail with `mark_processed requires a SQL or MongoDB connector`. |
| `table` | string | For SQL | — | Table to update. One of `table` or `collection` is required. |
| `collection` | string | For MongoDB | — | Collection to update. One of `table` or `collection` is required. |
| `key_column` | string | No | `id` | Column/field that identifies the row or document. |
| `key` | any | No | — | Key value. Omitted, the input item's `key_column` field is used. A missing key fails with `mark_processed requires key or key_column on input`. |
| `fields` | object | No | — | Columns/fields to write. |
| `set` | object | No | — | Alias for `fields`. Used only when `fields` is not set. |
| `unset` | string[] | No | — | MongoDB only: fields to remove from the document via `$unset`. |

## Example

```yaml
name: push_and_mark
connectors:
  source:
    type: sqlite
  destination:
    type: freshdesk
steps:
  - name: read_pending
    type: call_method
    config:
      connector: source
      method: select
      args:
        sql: SELECT * FROM contacts WHERE synced IS NULL
  - name: create_contact
    type: transform
    config: |
      {
        "id": input.id,
        "created": $callMethod({
          "connector": "destination",
          "method": "create",
          "resource": "api/v2/contacts",
          "data": { "name": input.name }
        })
      }
  - name: mark_synced
    type: mark_processed
    config: |
      {
        "connector": "source",
        "table": "contacts",
        "fields": {
          "synced": true,
          "destination_id": $string(input.created.id)
        }
      }
```

## Behavior notes

- **SQL Connectors** call the Connector's `update` method with `table`, `data` (the `fields`), and a filter of `{ key_column: key }`. When `fields` resolves to an empty object, no update is issued.
- **MongoDB Connectors** call `update_one` with `$set` built from `fields` and `$unset` built from `unset` (each listed field mapped to `''`). When both are empty, no update is issued.
- The key value defaults to the input item's `key_column` field (here `id`, carried through the transform). When a `call_method` step has replaced the item — for example with a create response that has no source key — pass `key` explicitly instead.
- The step returns the input item unchanged, so it can sit anywhere in the pipeline.

## Related

- [sync_record](/reference/steps/sync-record/) — combines the destination write with an equivalent writeback
- [filter](/reference/steps/filter/) — skip already-marked rows at the start of a delta pass
- [Migration playbook](/building/migration-playbook/) — delta-pass patterns that rely on marking