Import a dataset

Overview

H2O Hydrogen Torch lets you import your dataset using several supported data connectors.

note

H2O Hydrogen Torch requires your data to be formatted in a certain format base on the problem type the data aims to solve. To learn more, see Dataset formats.

Instructions

To import a dataset to H2O Hydrogen Torch, consider the following instructions:

1. Open the import dataset wizard

   In the H2O Hydrogen Torch navigation menu, click Import dataset.

2. Choose a data connector

   In the Source list, select the source (data connector) that you want to use. H2O Hydrogen Torch selects Demo Dataset by default.

   Demo Dataset

   1. In the File name list, select one of the Demo (preprocessed) datasets.

   AWS S3

   1. In the S3 bucket name box, enter the name of the S3 bucket name.

   2. In the AWS access key box, enter the AWS access key.

      Note

      You don't need to enter the AWS access key if the S3 bucket is public.

   3. In the AWS secret key box, enter the AWS secret key.

      Note

      You don't need to enter the AWS secret key if the S3 bucket is public.

   4. In the File name list, select the file you want to use.

   Google Cloud Storage

   1. In the GCS bucket name box, enter the name of the Google Cloud Storage bucket.

   2. In the GCS Service Account JSON box, enter the content of Google Cloud Service Account JSON file.

   3. In the File name list, select the file you want to use.

   Kaggle

   1. In the Kaggle API command box, enter a Kaggle API command.

   2. In the Kaggle username box, enter your username.

   3. In the Kaggle secret key box, enter your kaggle secret key.

   Azure Datalake

   1. In the Datalake connection string box, enter the Datalake connection string.

   2. In the Datalake container name box, enter the Datalake container name.

   3. In the File name box, enter the file name.

   H2O Drive

   1. In the File name list, select a dataset.

   Upload

   1. Click Browse.

      • Or drag and drop the file (dataset)

   2. Click Upload.

   3. Skip step 3.

   Local

   Note

   The Local source is unavailable on instances running in H2O AI Cloud.

   1. In the File location box, enter the path to the file you want to use. The box starts at your home directory.

   2. To browse instead, click a row in the directory table. H2O Hydrogen Torch opens the folder or selects the file you click, and updates the File location box with its path.

   3. To go up one level, click ...

3. Click Continue.

4. Define the import dataset settings

   Define the import dataset settings according to the dataset's problem type.

   Note

   • After importing a dataset through one of the supported data connectors for a problem type, H2O Hydrogen Torch automatically defines the dataset settings of the problem type by exploring the content of the imported dataset. Though, before saving the defined settings, you have the option to modify an erroneous value or option given to a dataset.
   • Before a successful dataset import, you need to define a few dataset settings before the dataset can be imported and used for an experiment. The required dataset settings depend upon the structure and content of the dataset; in other words, it depends on the problem type the dataset aims to solve. To learn about the particular required import dataset settings for a supported problem type, see Import dataset settings.

5. Click Continue.

6. Review the dataset preview

   Again, click Continue.

   Note

   Before you click Continue, please review the dataset preview.

Note

To learn how to edit the settings of a saved (imported) dataset, see Edit dataset.

---

Feedback

• Submit and view feedback for this page
• Send feedback about H2O Hydrogen Torch to cloud-feedback@h2o.ai