---
title: "Dataset creation and management"
slug: "aitems-dataset"
tags: ["AiTEMS", "Dataset"]
updated: 2026-09-17T09:33:52Z
published: 2026-09-17T09:33:52Z
canonical: "guide.ncloud-docs.com/aitems-dataset"
stale: true
---

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

# Dataset creation and management

Available in Classic and VPC

The Dataset creation and management section provides guidelines for configuring datasets and explains how to create, modify, upload, and delete datasets.

## Guidelines for configuring datasets

Since AiTEMS provides personalized recommendations based on dataset training, configuring datasets properly is crucial. Note the following when setting up the dataset.

- Dataset types include user, item, and interaction, and training requires all three dataset types. Below are descriptions of each dataset type:  

Descriptions for each dataset type:
  - user: Metadata containing user information (age, gender, etc.)
  - item: Metadata containing information related to an item (price, release date, category, etc.)
  - interaction: Metadata containing records of interactions between a user and an item
- Each dataset must include all required schema fields that correspond to its dataset type.

| Dataset type | Required fields (Not NULL) |
| --- | --- |
| user | USER_ID (string) |
| item | ITEM_ID (string) |
| interaction | USER_ID (string) ITEM_ID (string) TIMESTAMP (long) |
- Dataset fields must match the schema fields not only in field name but also in field order and letter case.
- Do not enter duplicate values in the required fields of the user and item datasets.

## Create a dataset

Create the datasets you will use for training. Datasets are managed by dataset name. When a dataset is created, it is assigned a unique dataset ID.

              Note

              

- To run training, you must create all dataset types.
- Before creating a dataset, prepare the dataset file you will upload. For details on structuring dataset files, see [Guidelines for configuring datasets](/docs/en/aitems-dataset#%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%85%8B%EA%B5%AC%EC%84%B1%EC%8B%9C%EC%A3%BC%EC%9D%98%EC%82%AC%ED%95%AD).
- Only csv or csv.gz files are supported.

To create a dataset:

1. Navigate to **Menu** > **All Services** > **AI Services** > **AiTEMS** in the NAVER Cloud Platform console.
2. Click the **Dataset** menu.
3. Click **[Create dataset]** button.
4. When the Create dataset interface appears, enter the dataset name and configure the dataset information.
  - **Dataset name**: Used as the identifier for managing the dataset. Enter 5–20 characters.
    - Must start with a letter and can include letters, numbers, _, and -.
  - **Description:**: Enter a description for the dataset.
  - **Dataset type:**: Select the dataset type (**user**/**item**/**interaction**) .
    - For details on dataset types, see [Guidelines for configuring datasets](/docs/en/aitems-dataset#%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%85%8B%EA%B5%AC%EC%84%B1%EC%8B%9C%EC%A3%BC%EC%9D%98%EC%82%AC%ED%95%AD).
    - You cannot change the **dataset type** after the dataset is created.
  - **Select schema**: Select an existing schema, or create a new one.
    - If no schema exists, select **Create schema** to create one.
    - When **Create schema** is selected, fields for configuring the new schema appear.
  - **Data selection**: Choose how to select the dataset file for training.
    - **Select from Object Storage**: Select this option when the dataset file has already been uploaded to the AiTEMS bucket in Object Storage.
    - **Select from file**: Select this option if the dataset file has not been uploaded to the AiTEMS bucket. When a file is selected and uploaded, it is automatically saved to the bucket.
5. If you selected **Create schema** under **Select schema**, configure the schema information and click **[Add]** button.
  - **Schema name**: Enter 3–20 characters. Used as the identifier for managing the schema.
  - **Schema description**: Enter a description for the schema.
  - **Field name**: Enter the field name exactly as it appears in the dataset file.
  - **Field type**: Select the data type to input (**string**/**float**/**long**/**double**/**int**/**boolean**/**null**)
  - **Categorical field**: Set to **Y** when entering data with predefined categories.
  - You can drag and drop ![i-aitems_order](https://cdn.document360.io/6998976f-9d95-4df8-b847-d375892b92c2/Images/Documentation/i-aitems_order.png?sv=2026-02-06&spr=https&st=2026-11-14T05%3A05%3A11Z&se=2026-11-14T06%3A05%3A11Z&sr=c&sp=r&sig=3%2FzxxK4J4DuXfN21%2BuYC8cG96M4alycuWlx9mtlzuDc%3D) to change the order of the fields.
6. If you seleted **Select from Object Storage** under **Data selection**, click the dataset file to upload.
  - Only files uploaded to the AiTEMS bucket can be selected.
7. If you selected **Select from file**, drag and drop the file into the **Drag files here or click to upload area,** or **click the area to select a file**.
  - The selected file is automatically uploaded to the bucket location displayed in **Bucket / Path**.
8. Click **[Create]** button.
9. In the notification popup, click **[OK]** button.
  - The dataset is created and added to the dataset list.

## Modify a dataset

You can update the dataset description, schema, and dataset file.

              Note

              

You cannot change the dataset type selected during creation.

To edit a dataset:

1. Navigate to **Menu** > **All Services** > **AI Services** > **AiTEMS** in the NAVER Cloud Platform console.
2. Click the **Dataset** menu.
3. In the dataset list, click the dataset you want to edit.
4. When the Edit dataset popup appears, apply the changes and click **[Save and upload]** button.
5. To change the dataset file, click **[Change path and upload]** button in the Data Edit popup, select the file you want to upload, and then click **[Change path and upload]** button again.
  - You can also change the dataset file by clicking **[Upload dataset]** button on the Dataset interface. (See [Dataset upload](/docs/en/aitems-dataset#%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%85%8B%EC%97%85%EB%A1%9C%EB%93%9C)).
6. In the notification popup, click **[OK]** button.
  - If the dataset file is changed, the status updates to **Pending**, and then updates again based on the upload result.

## Upload dataset

You can replace the dataset file.

              Note

              

- To replace a dataset file, the file must already be stored in the AiTEMS bucket in Object Storage.
- You can also replace the dataset file through [Edit dataset](/docs/en/aitems-dataset#%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%85%8B%EC%88%98%EC%A0%95).

To upload a dataset:

1. Navigate to **Menu** > **All Services** > **AI Services** > **AiTEMS** in the NAVER Cloud Platform console.
2. Click the **Dataset** menu.
3. In the dataset list, click the dataset for which you want to upload a dataset file.
4. Click the **[Upload dataset]** button.
5. In the Upload dataset popup, select the dataset file you want to upload and click **[Request upload]** button.
6. In the notification popup, click **[OK]** button.
  - The status changes to **Pending**, and will update based on whether the file upload succeeds.

## Deleting a dataset

To delete a dataset:

              Note

              

You cannot delete a dataset that is connected to a service. [Delete the relevant service](/docs/en/aitems-service#%EC%84%9C%EB%B9%84%EC%8A%A4%EC%82%AD%EC%A0%9C) or [change the dataset](/docs/en/aitems-train#%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%85%8B%EC%97%B0%EA%B2%B0) connected to the service before deleting the dataset.

1. Navigate to **Menu** > **All Services** > **AI Services** > **AiTEMS** in the NAVER Cloud Platform console.
2. Click the **Dataset** menu.
3. In the dataset list, click the dataset you want to delete.
4. Click **[Delete]** button.
5. When the Delete dataset popup appears, enter the dataset name and click **[Delete]** button.
  - The dataset is deleted and removed from the list.