Python & Pandas:无法删除列

Question

以下是我的部分数据：

USAF,   NCDC,  Date,     HrMn, I, Type,  QCP,  Dir, Q, I, Spd,   Q, 
034820,99999,19490801,0000,4,  SAO,    ,270,1,N,  5.7,1,
034820,99999,19490801,0100,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0200,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0300,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0400,4,  SAO,    ,270,1,N,  6.7,1,
034820,99999,19490801,0500,4,  SAO,    ,270,1,N,  8.7,1,
034820,99999,19490801,0600,4,  SAO,    ,270,1,N,  8.2,1,
034820,99999,19490801,0700,4,  SAO,    ,270,1,N,  8.2,1,

我尝试删除数据，但是它报告了一些列不存在。

ipath= "C:\Users\Administrator\Desktop\科研风速数据资料\Marham\\marham.txt"
uipath = unicode(ipath , "utf8")
file2 = open(uipath)
df = pd.read_csv(uipath,header=0)
# I can drop USAF, but unable to drop NCDC, I, Q, 
df.drop(['USAF', 'NCDC'], 1,inplace=True)
df.describe()

这些报告：

ValueError: labels ['NCDC'] not contained in axis

我的密码怎么了？我过去经常很好地使用df.drop(['a', 'b', 'c'], 1,inplace=True)。

更新：

df.columns.tolist()
['USAF',
 '   NCDC',
 '  Date',
 '     HrMn',
 ' I',
 ' Type',
 '  QCP',
 '  Dir',
 ' Q',
 ' I.1',
 ' Spd',
 '   Q',
 ' ']

Answer 1

read_csv的默认分隔符是逗号，但您的数据包含空格，这在标题和数据中引入了前导空格，如果您传递skipinitialspace=True，那么它导入的很好：

In [106]:
import io
import pandas as pd
t="""USAF,   NCDC,  Date,     HrMn, I, Type,  QCP,  Dir, Q, I, Spd,   Q, 
034820,99999,19490801,0000,4,  SAO,    ,270,1,N,  5.7,1,
034820,99999,19490801,0100,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0200,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0300,4,  SAO,    ,270,1,N,  4.6,1,
034820,99999,19490801,0400,4,  SAO,    ,270,1,N,  6.7,1,
034820,99999,19490801,0500,4,  SAO,    ,270,1,N,  8.7,1,
034820,99999,19490801,0600,4,  SAO,    ,270,1,N,  8.2,1,
034820,99999,19490801,0700,4,  SAO,    ,270,1,N,  8.2,1,"""
df = pd.read_csv(io.StringIO(t), skipinitialspace=True)
df

Out[106]:
    USAF   NCDC      Date  HrMn  I Type  QCP  Dir  Q I.1  Spd  Q.1  \
0  34820  99999  19490801     0  4  SAO  NaN  270  1   N  5.7    1   
1  34820  99999  19490801   100  4  SAO  NaN  270  1   N  4.6    1   
2  34820  99999  19490801   200  4  SAO  NaN  270  1   N  4.6    1   
3  34820  99999  19490801   300  4  SAO  NaN  270  1   N  4.6    1   
4  34820  99999  19490801   400  4  SAO  NaN  270  1   N  6.7    1   
5  34820  99999  19490801   500  4  SAO  NaN  270  1   N  8.7    1   
6  34820  99999  19490801   600  4  SAO  NaN  270  1   N  8.2    1   
7  34820  99999  19490801   700  4  SAO  NaN  270  1   N  8.2    1   

   Unnamed: 12  
0          NaN  
1          NaN  
2          NaN  
3          NaN  
4          NaN  
5          NaN  
6          NaN  
7          NaN

然后，您可以删除列：

In [108]:
df.drop(['USAF', 'NCDC'], 1,inplace=True)
df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 8 entries, 0 to 7
Data columns (total 11 columns):
Date           8 non-null int64
HrMn           8 non-null int64
I              8 non-null int64
Type           8 non-null object
QCP            0 non-null float64
Dir            8 non-null int64
Q              8 non-null int64
I.1            8 non-null object
Spd            8 non-null float64
Q.1            8 non-null int64
Unnamed: 12    0 non-null float64
dtypes: float64(3), int64(6), object(2)
memory usage: 768.0+ bytes