Question

我的问题与此one有关。我有一个名为“ test.csv”的文件，其中“ NA”是region的值。我想读为“ NA”而不是“ NaN”。但是，test.csv的其他列中缺少值，我想保留为“ NaN”。我怎样才能做到这一点？

# test.csv looks like this:

这是我尝试过的：

import pandas as pd
# This reads NA as NaN
df = pd.read_csv(test.csv)
df
    region  date    expenses
0   NaN   1/1/2019  53
1   EU    1/2/2019  NaN

# This reads NA as NA, but doesn't read missing expense as NaN
df = pd.read_csv('test.csv', keep_default_na=False, na_values='_')
df
    region  date    expenses
0   NA    1/1/2019  53
1   EU    1/2/2019  

# What I want:
    region  date    expenses
0   NA    1/1/2019  53
1   EU    1/2/2019  NaN

添加参数keep_default_na=False的问题是expenses的第二个值不能作为NaN读入。因此，如果我再尝试pd.isnull(df['value'][1])，则会返回为False。

Answer 1

对我来说，这可行：

df = pd.read_csv('file.csv', keep_default_na=False, na_values=[''])

给出：

  region      date  expenses
0     NA  1/1/2019      53.0
1     EU  1/2/2019       NaN

但是由于其他列中可能还有其他NaN，我宁愿安全起见，

df = pd.read_csv('file.csv')
df['region'] = df['region'].fillna('NA')

Answer 2

在指定keep_default=False时，所有默认值均不视为nan，因此您应指定它们：

使用keep_default_na=False, na_values= [‘’, ‘#N/A’, ‘#N/A N/A’, ‘#NA’, ‘-1.#IND’, ‘-1.#QNAN’, ‘-NaN’, ‘-nan’, ‘1.#IND’, ‘1.#QNAN’, ‘N/A’, ‘NULL’, ‘NaN’, ‘n/a’, ‘nan’, ‘null’]

Answer 3

这种方法对我有用：

import pandas as pd
df = pd.read_csv('Test.csv')
co1 col2  col3  col4
a   b    c  d   e
NaN NaN NaN NaN NaN
2   3   4   5   NaN

我复制了该值并创建了一个列表，默认情况下将其解释为 NaN，然后注释掉我想解释为非 NaN 的 NA。这种方法仍然将除 NA 之外的其他值视为 NaN。

#You can also create your own list of value that should be treated as NaN and 
# then pass the values to na_values and set keep_default_na=False.
        na_values = ["", 
                     "#N/A", 
                     "#N/A N/A", 
                     "#NA", 
                     "-1.#IND", 
                     "-1.#QNAN", 
                     "-NaN", 
                     "-nan", 
                     "1.#IND", 
                     "1.#QNAN", 
                     "<NA>", 
                     "N/A", 
        #              "NA", 
                     "NULL", 
                     "NaN", 
                     "n/a", 
                     "nan", 
                     "null"]
    
        df1 = pd.read_csv('Test.csv',na_values=na_values,keep_default_na=False )
    
              co1  col2  col3  col4
        a     b     c     d     e
        NaN  NA   NaN    NA   NaN
        2     3     4     5   NaN

阻止熊猫read_csv将NA解释为NaN，但保留NaN表示空值

3 个答案: