pandas 根据 A 字段分组取 B 字段最长的那一行

def test_for_longer():

    data = [{'foo': 'GD-20190101-010101', 'son': 'GD-123'},

            {'foo': 'GD-20190101-010101', 'son': 'GD-12345'},

            {'foo': 'GD-20190101-010102', 'son': 'GD-123'},

            {'foo': 'GD-20190101-010102', 'son': 'GD-123456'}]

    df = pd.DataFrame(data)

    df['size'] = df['son'].str.len()

    # df['size'] = df.apply(lambda x: len(x['son']), axis=1)

    idx = df.groupby('foo')['size'].idxmax()

    return df.loc[idx, ['foo', 'son']]

if __name__ == '__main__':

    val = test_for_longer()

    print(val)