Pandas to_csv method doesn't work well when use space separator

Code Sample, a copy-pastable example if possible

import numpy as np
import pandas as pd
s = pd.Series(np.random.randn(5), index=['a', 'b', 'c', 'd', 'e'])
s.to_csv('test.txt',sep=' ',float_format='%12.5f')

Problem description

The contents of test.txt is

a "        0.09048"
b "        0.13346"
c "        0.13251"
d "        0.00243"
e "        0.03104"

not sure where is the " from

Expected Output

same result without "

Output of pd.show_versions()

python: 3.5.0.final.0 OS:Linux pandas: 0.20.3

Comment From: TomAugspurger

Hmm I think everything is working properly here. To make things easier to explain, we'll do print(s.to_frame().to_csv(...), since Series.to_csv doesn't have all the options DataFrame.to_csv does yet (we have another issue for that).

With the options you gave, we need the quotes, else we generate an invalid CSV, since the header row would have just a single field and the rest of the rows would have many (the number of spaces).

In [24]: print(s.to_frame().to_csv(sep=" ", float_format='%12.5f'))
 0
a "     0.37333"
b "     1.72104"
c "     0.01158"
d "    -0.38504"
e "     0.81918"

To say you want no quoting no matter what, use csv.QUOTE_NONE

In [32]: print(s.to_frame().to_csv(sep=' ',float_format='%12.5f', quoting=csv.QUOTE_NONE))
...
Error: need to escape, but no escapechar set

And it's complaining since you'd generate an invalid CSV that way, so you need an escapechar:

In [33]: print(s.to_frame().to_csv(sep=' ',float_format='%12.5f', quoting=csv.QUOTE_NONE, escapechar='\\'))
 0
a \ \ \ \ \ 0.37333
b \ \ \ \ \ 1.72104
c \ \ \ \ \ 0.01158
d \ \ \ \ -0.38504
e \ \ \ \ \ 0.81918

Comment From: tonymonk007

Now I am clear. Thanks a lot.