为什么在pd.read_csv（）之后创建虚拟“ 0”行？

Question

Why there is an additional row named '0' after read_csv?

In the following code, I saved df1 to csv file and read it back. However, there is an additional row named '0'. How can I avoid it?

d1 = {'a' : 1, 'b' : 2, 'c' : 3}
df1=pd.Series(d1)
print('\ndf1:'); print(df1)

> df1:
> a    1
> b    2
> c    3
> dtype: int64

df1.to_csv("df1.csv")
df1=pd.read_csv("df1.csv",  index_col=0, header=None)
print('\ndf1:'); print(df1)

> df1:
>   1
> 0    <<<<---- ????
> a  1
> b  2
> c  3

Answer 1

That is not a dummy row - that is the name of your indexing Column.

You can check it by running:

> df.index.name
0

You can change it by setting it too:

> df.index.name = "my_index"

          1
my_index   
a         1
b         2
c         3

Answer 2

One way to avoid it is to create a DataFrame (as your variable name suggests) instead of a Series (which you have atm).

import pandas as pd
d1 = {'a' : 1, 'b' : 2, 'c' : 3}
df1 = pd.Series(d1).to_frame()    # Use to_frame() here
df1.to_csv("df1.csv")
df1 = pd.read_csv("df1.csv", index_col=0)

print(df1) gives:

   0
a  1
b  2
c  3

---

Another solution would be to actually specify names when you import. In case you opt for this solution I would however rename the df1 to s1 when you create a series for readability.

df1 = pd.read_csv("df1.csv", names=['index','values'], index_col='index')

And you get:

       values
index        
a           1
b           2
c           3

Full example:

d1 = {'a' : 1, 'b' : 2, 'c' : 3}
s1 = pd.Series(d1)
print('s1:\n{}\n'.format(s1))

s1.to_csv("df1.csv")
df1 = pd.read_csv("df1.csv", names=['index','values'], index_col='index')
print('df1:\n{}\n'.format(df1))