pandas: dataframe to_csv, how to set column names

Question

Code snippet:

import numpy as np
import pandas as pd
myseries=pd.Series(np.random.randn(5))
df=pd.DataFrame(myseries)
df.to_csv("output.csv")

Output:

      0
0    0.51..
1    0.14..
2    -0.68..
3    0.48..
4    1.89..

I want the column name to be "values" instead of 0. How can I do that?
I suppose that I should replace the last statement with df.to_csv("output.csv", columns = ["values"]). But I got the keyerror: u"None of [['values']] are in the [columns]" I don't know what that means.

[update]
Many answers say that I should use df.columns=['values']. Well, that does not work for me. I concern not only what the dataframe is like, but also what the csv file is like. The dataframe looks all right, but the csv file is not. That's the confusing part.

...
df.columns=["values"]
df.to_csv("output.csv")

It says:IOError: [Errno 13] Permission denied: 'output.csv'.
I then use the absolute path "C:\Users\myname\Desktop\output.csv", the error is like: IOError: [Errno 13] Permission denied: 'C:\\Users\\myname\\Desktop\\output.csv' I don't know why this error, but this is rather confusing.
For further information, I installed anaconda-2.7 on win10. I tested the code with spyder.

Answer 1

You can set column name in DataFrame constructor:

df = pd.DataFrame(myseries, columns=['values'])
df.to_csv("output.csv")

Or:

df = pd.DataFrame({'values':myseries})
print (df)
     values
0 -0.429758
1 -0.019931
2  1.189596
3  1.309223
4 -0.337061

df.to_csv("output.csv")

Or set parameter header in DataFrame.to_csv:

df = pd.DataFrame(myseries)
df.to_csv("output.csv", header=['values'])

Or in Series.to_csv:

myseries.to_csv("output.csv", header=['values'])