TypeError：不理解数据类型“列表”

Question

I have a Series object, returned by pandas groupby , which has elements of numpy.ndarray type. I would like to convert ndarrays to lists, preferably without using loops. I tried to use pandas.Series.astype but I got error: TypeError: data type 'list' not understood . Why is that when documentation says that

Use a numpy.dtype or Python type to cast entire pandas object to the same type.

and list is Python buil-id data type.

Example:

a = {'Id': {0: 1, 1: 2, 2: 3},
 'col': {0: 5.1, 1: 4.9, 2: 4.9}}
d = pd.DataFrame.from_dict(a)
attr_unique_val = d.groupby('col')['Id'].unique()
attr_unique_val = attr_unique_val.astype('list')

Answer 1

You can just use apply and lambda, using the example you provided:

    a = {'Id': {0: 1, 1: 2, 2: 3},
    'col': {0: 5.1, 1: 4.9, 2: 4.9}}
    d = pd.DataFrame.from_dict(a)
    attr_unique_val = d.groupby('col')['Id'].unique()
    attr_unique_val = attr_unique_val.apply(lambda x: x.tolist())

Answer 2

If you want to convert ndarrays to list then you can use tolist() method. For example- the ndarray name is 'narr' then you can write narr.tolist() .