Error in rpy2- No (non-missing) observations

I m working with rpy2 on Ubuntu 1104. I m following the COX file. The steps given in file are in R. I have to do same steps in python using rpy2. I didn't get any tutorial with rpy2. I managed to write following,

from rpy2.robjects.packages import importr
from rpy2.robjects import IntVector, Formula
import rpy2.robjects as ro
cox = importr("survival")
csv = ro.vectors.DataFrame.from_csvfile('Rossi.txt', header=True, sep=' ')
fmla = Formula('Surv(week, arrest) ~ fin + age + race + wexp + mar + paro + prio')
mod_aalison = cox.coxph (fmla, data=csv)

But I'm getting following error,

>>> mod_aalison = cox.coxph (fmla, data=csv)
Error in function (formula, data, weigh开发者_运维百科ts, subset, na.action, init, control,  : 
  No (non-missing) observations
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.7/dist-packages/rpy2/robjects/functions.py", line 83, in __call__
    return super(SignatureTranslatedFunction, self).__call__(*args, **kwargs)
  File "/usr/lib/python2.7/dist-packages/rpy2/robjects/functions.py", line 35, in __call__
    res = super(Function, self).__call__(*new_args, **new_kwargs)
rpy2.rinterface.RRuntimeError: Error in function (formula, data, weights, subset, na.action, init, control,  : 
  No (non-missing) observations

Am I missing any thing? I have no prior experience on R. I'm not sure the if data in function coxph has to be in dataframe format. Any help is appreciated.

This is link to Rossi.txt

---

You read the data using read.csv with sep set to a single space, while the file has two spaces. This makes R interpret this as an existence of empty columns, which then mess with column names and finally create error in coxph.

Read the data with read.table, and it should be ok.