> ## Documentation Index
> Fetch the complete documentation index at: https://docs.auxia.io/llms.txt
> Use this file to discover all available pages before exploring further.

# CSV Field Reference

This page provides detailed information about each column in the bulk upload CSV template.

## 1. Base Fields

* **External treatment ID**
  * **Type**: String
  * **Required**: No
  * **Description**: External identifier for integration with other systems
  * **Notes**:
    * Only available if enabled in your project settings
    * Must be unique within your project if provided

* **Tags**

  * **Type**: String
  * **Required**: No
  * **Description**: Tags for organizing and categorizing treatments
  * **Format**: Use semicolons (`;`) to separate multiple tags
  * **Formatting Rules**
    * Use semicolons (`;`) to separate tags
    * Do not include spaces after semicolons
    * Do not leave empty values between semicolons
    * Tags are case-sensitive

  <Note>
    All tags must already exist in your project before using them in bulk upload. Create tags in the console first if needed.
  </Note>

* **Treatment type**\*
  * **Type**: String
  * **Required**: Yes
  * **Description**: The treatment type for this treatment
  * **Example**: `POPUP_BANNER`, `PUSH_NOTIFICATION`
  * **Notes**:
    * The Treatment Type should already exist
    * All treatments in a CSV must have the same Treatment type

* **Target action**
  * **Type**: String
  * **Format**:
    * **Basic Event**
      * For events without specific property requirements.
      * **Syntax**: `EventName`
      * **Example**:  `purchase`
    * **Event with Properties**
      * For events that require specific property values.
      * **Syntax**: `EventName (PropertyName1: PropertyValue1 ; PropertyName2: PropertyValue2)`
      * **Single Property Example**: `purchase (product_type: premium)`
      * **Multiple Properties Example**: `level_end (level_number: 5 ; result: success ; difficulty: hard)`
  * **Important Notes**
    * Event name must exist in your project
    * Property names must exactly match the event's property definitions
    * Property values must exactly match existing values for that property
    * All specified properties must exist for the event

* **Name**\*
  * **Type**: String
  * **Required**: Yes
  * **Description**: The treatment name displayed in the console

* **Description**\*
  * **Type**: String
  * **Required**: Yes
  * **Description**: Detailed description of the treatment's purpose

* **Surfaces**\*
  * **Type**: String (semicolon-separated)
  * **Required**: Yes
  * **Description**: Surfaces to include for this treatment
  * **Format**: Use semicolons (`;`) to separate multiple surfaces
  * **Notes**
    * Only surfaces associated with the treatment type can be specified
    * At least one surface must be specified
    * Surface names must exactly match existing surfaces

* **MEG**
  * **Type**: String
  * **Required**: No
  * **Description**: Mutual Exclusion Group name
  * **Example**: `holiday_promotions`

<Note>
  The Mutual Exclusion Group (MEG) must already exist in your project before using it in bulk upload. Create the MEG in the console first if needed.
</Note>

***

## 2. Launch Schedule

* **Start timestamp**\*
  * **Type**: Timestamp (UTC)
  * **Required**: Yes
  * **Description**: When the treatment becomes active
  * **Format**: `MM/dd/yyyy hh:mm a`
  * **Example**: `06/01/2024 12:00 AM`
  * **Notes**:
    * Must be in UTC timezone
    * Use 12-hour format with AM/PM
    * Month and day should be zero-padded (e.g., `01` not `1`)

* **End timestamp**
  * **Type**: Timestamp (UTC)
  * **Required**: No
  * **Description**: When the treatment expires
  * **Format**: `MM/dd/yyyy hh:mm a`
  * **Example**: `12/31/2024 11:59 PM`
  * **Notes**:
    * Must be in UTC timezone
    * Use 12-hour format with AM/PM
    * If not provided, treatment runs indefinitely

***

## 3. Content Fields

Content fields vary based on your treatment type. They fall into two categories:

* **Language-Agnostic Fields**

  Fields that apply to all languages.

  * **Column Name**: `field_name`
  * **Example**: `image_url`, `deep_link`

* **Language-Specific Fields**

  Fields that have different content for each language.

  * **Column Name**: `field_name_languagecode`
  * **Examples**:
    * `title_en`, `title_ja`
    * `body_en`, `body_ja`
    * `cta_text_en`, `cta_text_ja`

  **Notes**:

  * Each field has a maximum character limit (varies by treatment type and field)
  * Exceeding the character limit will cause validation errors

***

## 4. Qualification Criteria

Qualification criteria determine which users are eligible to receive the treatment. Each qualification criterion is defined by four columns.

* **Column Pattern**

  For each qualification criterion number N (where N = 1, 2, 3, ...):

  | Column                                             | Description                                          |
  | -------------------------------------------------- | ---------------------------------------------------- |
  | `Qualification field N`                            | The data field to evaluate                           |
  | `Qualification condition N`                        | The comparison condition                             |
  | `Qualification value N`                            | The value to compare against                         |
  | `Apply qualification N if attribute not available` | Behavior when data is missing (User Attributes only) |

* **Qualification field N**
  * **Description**: The data field to check for eligibility. Can be one of two types:
  * **Types**:
    * **Regular Data Field (by ID)**
      * Use the numeric data field ID from your project
      * **Example**: `12345`
    * **Self Referencing Treatment Interaction**
      * **What it is**: A special data field that tracks how users have interacted with the treatment you're currently creating
      * **Format**: `InteractionType_(THIS_TREATMENT)_TimeWindow`
      * **Note**: `(THIS_TREATMENT)` is a placeholder that automatically refers to the treatment being created
      * **Example**: Limit how often a treatment is shown to users
        * **In CSV, enter**: `Sent_(THIS_TREATMENT)_last_1d` with condition `is less than` and value `2`
        * **In console, displays as**: `auxia_treatment_sent treatment_id:12345 last_1d<2` (where 12345 is the treatment ID)
      * **Supported Interaction Types**:
        * `Sent` - Treatment was sent to user
        * `Viewed` - User viewed the treatment
        * `Clicked` - User clicked on the treatment
        * `Snoozed` - User snoozed the treatment
        * `Dismissed` - User dismissed the treatment
        * `Cta_completed` - User completed the call-to-action
        * `Swipe` - User swiped on the treatment
      * **Supported Time Windows**:
        | Time Window        | Returns   | Description                             |
        | ------------------ | --------- | --------------------------------------- |
        | `first_timestamp`  | TIMESTAMP | When interaction first occurred         |
        | `latest_timestamp` | TIMESTAMP | When interaction most recently occurred |
        | `last_1d`          | INTEGER   | Count in last 1 day                     |
        | `last_7d`          | INTEGER   | Count in last 7 days                    |
        | `last_14d`         | INTEGER   | Count in last 14 days                   |
        | `last_28d`         | INTEGER   | Count in last 28 days                   |
        | `last_56d`         | INTEGER   | Count in last 56 days                   |
        | `days_since_first` | INTEGER   | Days since first occurrence             |
        | `days_since_last`  | INTEGER   | Days since last occurrence              |

* **Qualification condition N**
  * **Description**: The comparison operator to use. Valid conditions depend on the data field type.
  * **Conditions by Type**:
    * **For INTEGER and DOUBLE**:

      * `is` - Equals
      * `is not` - Not equals
      * `is greater than` - Greater than
      * `is at least` - Greater than or equal to
      * `is less than` - Less than
      * `is at most` - Less than or equal to
      * `is any of` - Matches any value in a list
      * `is not any of` - Does not match any value in a list

      <Note>
        The `between` condition is not supported. Instead, use two separate qualification criteria: one with `is at least` for the minimum value and another with `is at most` for the maximum value.
      </Note>

    * **For STRING**:
      * `is` - Exactly equals
      * `is not` - Not equals
      * `contains` - Contains substring
      * `does not contain` - Does not contain substring
      * `is any of` - Matches any value in a list
      * `is not any of` - Does not match any value in a list

    * **For BOOLEAN**:
      * `is` - Equals (true or false)

    * **For TIMESTAMP\_SECONDS**:
      * `at or before` - At or before timestamp
      * `at or after` - At or after timestamp
      * `is older than` - Is before current time minus a duration
      * `> current time minus` - Is after current time minus a duration
      * `has happened` - Checks if the event has occurred (value: true) or not occurred (value: false)

    * **For INTEGER\_LIST, DOUBLE\_LIST, STRING\_LIST**:
      * `contains` - List contains value
      * `does not contain` - List does not contain value
      * `contains any of` - List contains any value in specified list
      * `does not contain any of` - List does not contain any value in specified list

* **Qualification value N**
  * **Description**: The value to compare against. Format depends on the condition and data field type.

* **Apply qualification N if attribute not available**
  * **Type**: String (Boolean)
  * **Required**: No (only applicable for User Attribute data fields)
  * **Description**: Controls what happens when a User Attribute data field has missing or unavailable data
  * **Valid Values**:
    * `TRUE` - If the attribute is missing, treat the qualification as satisfied
    * `FALSE` - If the attribute is missing, treat the qualification as not satisfied
    * Empty - Default behavior
  * **Important Notes**:
    * This field only applies to User Attribute data fields (data fields based on user properties)
    * Cannot be set for Self Referencing Treatment Interaction fields or Event-based fields
    * If you try to set this value for non-User Attribute fields, validation will fail
  * **Example Scenario**:
    If you're targeting users with `country` = `US`, but some users don't have a country value:
    * Set to `FALSE`: Users without a country value won't receive the treatment even if other criteria are satisfied
    * Set to `TRUE`: Users without a country value will be treated given that other criteria are satisfied

***

**Next Step**: Once your CSV is filled out, proceed to [Upload and Validate](/bulk-upload/upload-and-validate) for validation rules, error troubleshooting, and to upload your file.