Load a specific field in load_dataset

[SubhashiniRajesh26]

Is it possible to load only a specific field of json file in load_datset()? Suggestions are most welcome

[bmaxdk]

Yes, simply you can load_dataset('json', data_files='use_your_json_file.json', field='data')

ex:

{
    "data": [
      {
        "id": "1",
        "text": "I love programming.",
        "label": "positive"
      },
      {
        "id": "2",
        "text": "I don'like bugs.",
        "label": "negative"
      }
    ],
    "metadata": {
      "version": "1.0",
      "source": "generated"
    }
  }
  

from datasets import load_dataset

# If you want to load the data field from the JSON file then
dataset = load_dataset('json', data_files='classification.json', field='data')

print(dataset)