pandas——如何处理NaN数据

函数	含义
dropna(axis, how)	删除整行或整列（axis是按照啥方向， how是删的条件，这俩都是可选的）
d.fillna(num)	用num来替代其中的NaN
d.isnull()	返回一个bool matrix，NaN表示为True
np.any(d.isnull()) == True	判断d中有没有NaN

axis挺迷的：
知乎的

dropna(self, axis, how)

import pandas as pd
import numpy as np

d = pd.DataFrame(np.linspace(2, 17, 16).reshape(4, 4), ['a', 'b', 'c', 'd'], ['A', 2, 3, 4])
d.loc['b', 3] = np.nan
d.loc['c', 2] = np.nan
print(type(d))  # <class 'pandas.core.indexes.datetimes.DatetimeIndex'>
print(d)
print(d.dropna())  #

print(d.dropna(axis=1))

how：

• any：表示只要有NaN就删
• all：表示全都是NaN才删

print(d.dropna(how='all'))

d.fillna(num)

import pandas as pd
import numpy as np

d = pd.DataFrame(np.linspace(2, 17, 16).reshape(4, 4), ['a', 'b', 'c', 'd'], ['A', 2, 3, 4])
d.loc['b', 3] = np.nan
d.loc['c', 2] = np.nan
print(type(d))  # <class 'pandas.core.indexes.datetimes.DatetimeIndex'>
print(d)
print(d.fillna(-1))

d.isnull()

import pandas as pd
import numpy as np

d = pd.DataFrame(np.linspace(2, 17, 16).reshape(4, 4), ['a', 'b', 'c', 'd'], ['A', 2, 3, 4])
d.loc['b', 3] = np.nan
d.loc['c', 2] = np.nan
print(type(d))  # <class 'pandas.core.indexes.datetimes.DatetimeIndex'>
print(d)
print(d.isnull())

np.any(d.isnull()) == True

import pandas as pd
import numpy as np

d = pd.DataFrame(np.linspace(2, 17, 16).reshape(4, 4), ['a', 'b', 'c', 'd'], ['A', 2, 3, 4])
d.loc['b', 3] = np.nan
d.loc['c', 2] = np.nan
print(type(d))  # <class 'pandas.core.indexes.datetimes.DatetimeIndex'>
print(d)
print(np.any(d.isnull()) == True)