from sklearn.datasets import fetch_20newsgroups
import pandas as pd

def twenty_newsgroup_to_csv():
    newsgroups_train = fetch_20newsgroups(subset='train')

    df = pd.DataFrame([newsgroups_train.data, newsgroups_train.target.tolist()]).T
    df.columns = ['text', 'target']

    targets = pd.DataFrame( newsgroups_train.target_names)
    targets.columns=['title']

    out = pd.merge(df, targets, left_on='target', right_index=True)
    out['date'] = pd.to_datetime('now')
    out.to_csv('20_newsgroup.csv')

twenty_newsgroup_to_csv()