Construct
Series can be created using the following constructor
# data : 数据源,ndarray、list、dic、常量等
# index : 索引,唯一和散列等,与数据的长度相同
# dtype : 指定数据类型,默认系统推断数据类型
# copy : 复制数据,默认为 Flase
pd.Series( data, index, dtype, copy)
- Create empty series
s = pd.Series()
print(s)
'''Series([], dtype: float64)'''
- via ndarray
s = pd.Series(np.array(['a','b','c']))
print(s)
'''
0 a
1 b
2 c
dtype: object'''
- via list
s = pd.Series(['1','a','3',2])
print(s)
'''
0 1
1 a
2 3
3 2
dtype: object'''
- via dic dictionary
# 字典的 key 即为标签
dic = {'name':'luo', 'age':25, 'sex':'F'}
s = pd.Series(dic)
print(s)
'''
age 25
name luo
sex F
dtype: object'''
- by scalar
s = pd.Series(2, index=[0, 1, 2])
print(s)
'''
0 2
1 2
2 2
dtype: int64'''
Designated subscript
s = pd.Series(np.arange(3), index=list('ABC'))
print(s)
'''
A 0
B 1
C 2
dtype: int64'''
retrieve data
The data takes s in the above example as an example
- by index
data = s[1]
print('通过索引取值:{}'.format(data))
'''通过索引取值:1'''
If the series has a label index, we can also get the value by the label
- by label
a. Array
data = s[['B','A']]
print('通过标签取值:\n{}'.format(data))
'''
通过标签取值:
B 1
A 0
dtype: int64'''
b. Sliced
data = s['A':'C']
print('通过标签取值:\n{}'.format(data))
'''
A 0
B 1
C 2
dtype: int64'''