panda缺失值处理

http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.fillna.html

import pandas as pd
import numpy as np

df=pd.DataFrame(np.random.rand(5,6))
df

	0	1	2	3	4	5
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	0.710527	0.361684	0.094026	0.895336	0.848446
2	0.552851	0.456602	0.410653	0.835223	0.769456	0.803724
3	0.392067	0.397841	0.393220	0.745361	0.360859	0.383625
4	0.820769	0.893594	0.312887	0.378115	0.584261	0.214013

# Make a few areas have NaN values
df.iloc[1:3,1] = np.nan
df.iloc[3:,3] = np.nan
df.iloc[2,5]=np.nan
df

	0	1	2	3	4	5
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	NaN	0.361684	0.094026	0.895336	0.848446
2	0.552851	NaN	0.410653	0.835223	0.769456	NaN
3	0.392067	0.397841	0.393220	NaN	0.360859	0.383625
4	0.820769	0.893594	0.312887	NaN	0.584261	0.214013

df.isnull()

	0	1	2	3	4	5
0	False	False	False	False	False	False
1	False	True	False	False	False	False
2	False	True	False	False	False	True
3	False	False	False	True	False	False
4	False	False	False	True	False	False

#显示存在缺失值的行列，清楚的确定缺失值的位置
df[df.isnull().values==True]

.dataframe thead tr:only-child th { text-align: right; } .dataframe thead th { text-align: left; } .dataframe tbody tr th { vertical-align: top; }

	0	1	2	3	4	5
1	0.200017	NaN	0.361684	0.094026	0.895336	0.848446
2	0.552851	NaN	0.410653	0.835223	0.769456	NaN
2	0.552851	NaN	0.410653	0.835223	0.769456	NaN
3	0.392067	0.397841	0.393220	NaN	0.360859	0.383625
4	0.820769	0.893594	0.312887	NaN	0.584261	0.214013

#填充缺失数据
df.fillna(0)

.dataframe thead tr:only-child th { text-align: right; } .dataframe thead th { text-align: left; } .dataframe tbody tr th { vertical-align: top; }

	0	1	2	3	4	5
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	0.000000	0.361684	0.094026	0.895336	0.848446
2	0.552851	0.000000	0.410653	0.835223	0.769456	0.000000
3	0.392067	0.397841	0.393220	0.000000	0.360859	0.383625
4	0.820769	0.893594	0.312887	0.000000	0.584261	0.214013

#根据(axis=0列,axis=1行)的前一个值进行填充,method : {‘backfill’, ‘bfill’, ‘pad’, ‘ffill’, None}, default None
df.fillna(axis=0,method='ffill')

	A	B	C	D	E	F
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	0.546635	0.361684	0.094026	0.895336	0.848446
2	0.552851	0.546635	0.410653	0.835223	0.769456	0.848446
3	0.392067	0.397841	0.393220	0.835223	0.360859	0.383625
4	0.820769	0.893594	0.312887	0.835223	0.584261	0.214013

df.columns=['A','B','C','D','E','F']
df

	A	B	C	D	E	F
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	NaN	0.361684	0.094026	0.895336	0.848446
2	0.552851	NaN	0.410653	0.835223	0.769456	NaN
3	0.392067	0.397841	0.393220	NaN	0.360859	0.383625
4	0.820769	0.893594	0.312887	NaN	0.584261	0.214013

#Replace all NaN elements in column ‘A’, ‘B’, ‘C’, and so on with 0, 1, 2, and 3 respectively.
values = {'A': 0, 'B': 1, 'C': 2, 'D': 3,'E':4,'F':5}
df.fillna(value=values)

	A	B	C	D	E	F
0	0.399916	0.546635	0.729255	0.992735	0.747917	0.157603
1	0.200017	1.000000	0.361684	0.094026	0.895336	0.848446
2	0.552851	1.000000	0.410653	0.835223	0.769456	5.000000
3	0.392067	0.397841	0.393220	3.000000	0.360859	0.383625
4	0.820769	0.893594	0.312887	3.000000	0.584261	0.214013