# User identifiers association

This document import allows you to merge UserPoint by associating their [user identifiers](https://developer.mediarithmics.io/user-points#user-identifiers). Each line in the document represents a different user identifiers association 

{% hint style="info" %}
This is only supported for datamarts using a UserPoint system version of `v201812` or later. 
{% endhint %}

## How-to

1. Use the [bulk import ](https://developer.mediarithmics.io/data-streams/data-ingestion/bulk-processing)endpoints to create a [document import](https://developer.mediarithmics.io/data-streams/data-ingestion/bulk-processing/..#document-import) with the`USER_IDENTIFIERS_ASSOCIATION_DECLARATIONS`document type and `APPLICATION_X_NDJSON` mime type. Only `ndjson` data is supported for user activities.
2. Create an[ execution ](https://developer.mediarithmics.io/data-streams/data-ingestion/bulk-processing/..#create-an-execution)with your commands formatted in `ndjson` .

### User identifiers association command

Each line will create/merge a UserPoint that has all the specified identifiers

| field       | type                      | description                                      |
| ----------- | ------------------------- | ------------------------------------------------ |
| identifiers | UserIdentifierResource\[] | An array of User Identifier Resource of any type |

User identifier resource can be of three shapes. Either email or user agent or user account id. They correspond with the different types of [user identifiers](https://developer.mediarithmics.io/user-points#user-identifiers).

#### Email

| field | type              | description                                                              |
| ----- | ----------------- | ------------------------------------------------------------------------ |
| type  | "USER\_EMAIL"     | The type of the identifier.                                              |
| hash  | String            | A hash of the email. The hashing function should be unique per datamart. |
| email | String (optional) | the email address                                                        |

#### User Agent

| field           | type          | description                 |
| --------------- | ------------- | --------------------------- |
| type            | "USER\_AGENT" | The type of the identifier. |
| user\_agent\_id | String        | The user agent ID           |

#### User Account

| field             | type              | description                                                                                    |
| ----------------- | ----------------- | ---------------------------------------------------------------------------------------------- |
| type              | "USER\_ACCOUNT"   | The type of the identifier.                                                                    |
| user\_account\_id | String            | The User Account ID                                                                            |
| compartment\_id   | String (optional) | The Compartment ID. If you don't input the compartment id it will fall back on the default one |

### Example

```bash
# Create the document import
curl -X POST \
  https://api.mediarithmics.com/v1/datamarts/<DATAMART_ID>/document_imports \
  -H 'Authorization: <YOUR_API_TOKEN>' \
  -H 'Content-Type: application/json' \
  -d '{
	"document_type": "USER_IDENTIFIERS_ASSOCIATION_DECLARATIONS",
	"mime_type": "APPLICATION_X_NDJSON",
	"encoding": "utf-8",
	"name": "<YOUR_DOCUMENT_IMPORT_NAME>"
}'
```

```bash
# Create the execution
curl -X POST \
  https://api.mediarithmics.com/v1/datamarts/1162/document_imports/<DOCUMENT_IMPORT_ID>/executions \
  -H 'Authorization: <API_TOKEN>' \
  -H 'Content-Type: application/x-ndjson' \
  -d '
    { 
      "identifiers":[
        { "type": "USER_EMAIL", "hash":"<EMAIL_HASH>" }, 
        { "type": "USER_AGENT", "user_agent_id": "<USER_AGENT_ID>" }, 
        { "type": "USER_ACCOUNT", "user_account_id": "<USER_ACCOUNT_ID>",  "compartment_id": "<COMPARTMENT_ID>" }
      ]
    }
  '
```

{% hint style="success" %}
You can, of course, add different identifier types at the same time. Please note that the uploaded data is in `ndjson` and not `json`. That means the different additions are not separated by commas, but by a line separator `\n`
{% endhint %}