Pandas DataFrame serialization

Question

I'm having trouble writing the entries of a pandas dataframe to a stringbuffer.

It's possible to initialize a dataframe by passing a stringbuffer to the read_csv function.

In [80]: buf = StringIO('a,b\n1,2\n')
In [81]: df = pandas.read_csv(buf)
In [82]: df
Out[82]: 
   a  b
0  1  2

To do the opposite is not straight forward as the DataFrame.to_csv function only accepts a string file path.

Is there any good reason for this behaviour? What's the best way to serialize a pandas DataFrame without storing the contents on disk first?

Wes McKinney · Accepted Answer · 2012-02-08 23:15:46Z

3

This is pretty much just an oversight / inconsistency. I created a GitHub issue here for it:

EDIT: Implemented this today so you can pass a StringIO to to_csv, in git master now and will be part of forthcoming 0.7.0 release

answered Feb 8, 2012 at 17:10

Wes McKinney

106k32 gold badges146 silver badges109 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

link to issue is broken