>>> obj = Series(['c','a','d','a','a','b','b','c','c'])

>>> obj.unique()

array(['c', 'a', 'd', 'b'], dtype=object)

>>> obj

0    c

1    a

2    d

3    a

4    a

5    b

6    b

7    c

8    c

dtype: object

>>> type(obj.unique())

<class 'numpy.ndarray'>#注意这里返回的不再是Series对象，而是ndarray的一维数组

>>> new_array = obj.unique()

>>> new_array

array(['c', 'a', 'd', 'b'], dtype=object)

>>> new_array.sort()

>>> new_array

array(['a', 'b', 'c', 'd'], dtype=object)

>>> import numpy as np

>>> new_array = obj.unique()

>>> new_array

array(['c', 'a', 'd', 'b'], dtype=object)

>>> na = np.sort(new_array)

>>> na

array(['a', 'b', 'c', 'd'], dtype=object)

>>> obj

0    c

1    a

2    d

3    a

4    a

5    b

6    b

7    c

8    c

dtype: object

>>> obj_c= obj.value_counts()

>>> obj_c

c    3

a    3

b    2

d    1

dtype: int64

>>> pd.value_counts(obj)#默认是降序

c    3

a    3

b    2

d    1

dtype: int64

>>> pd.value_counts(obj,sort =False)#对统计结果不排序

a    3

b    2

d    1

c    3

dtype: int64

>>> mask = obj.isin(['a','c'])

>>> mask

0     True

1     True

2    False

3     True

4     True

5    False

6    False

7     True

8     True

dtype: bool

>>> obj[mask]

0    c

1    a

3    a

4    a

7    c

8    c

dtype: object