rpy2: prevent StrVector from factorisation when put into a DataFrame

Question

In rpy2, I noticed that a StrVector is factorised once it's put into a DataFrame. An example is the following.

import rpy2.robjects as ro

series_1 = ("0", "0", "0", "0")
series_1_robject = ro.StrVector(series_1)  # => ['0', '0', '0', '0']
df = ro.DataFrame({"series_1": series_1_robject})    # => FactorVector [1, 1, 1, 1]

And...

>>> df[0][1]
1

It appears that when I build a DataFrame, my nice StrVector gets factorised, so 0 corresponds to factor value 1 (being R is 1-indexed), and so on. But how would I stop this from happening? It's pretty essential for me that when the input vector (series_1) is 0,0,0...,0, its representation in the resulting DataFrame will be 0, not 1. So far, I have not really been able to find anything on this matter in the documentation....

Answer 1

As per the note here, you can prevent this conversion to FactorVector by wrapping the StrVector with a call to ro.r.I() (the "as-is" function in R):

In [1]: import rpy2.robjects as ro

In [2]: series_1 = ("0", "0", "0", "0")

In [3]: series_1_robject = ro.StrVector(series_1)

In [4]: df = ro.DataFrame({"series_1": series_1_robject})

In [5]: df.rx2("series_1")
Out[5]:
R object with classes: ('factor',) mapped to:
<FactorVector - Python:0x113a39368 / R:0x7f8d15882e40>
[       1,        1,        1,        1]

In [6]: df = ro.DataFrame({"series_1": ro.r.I(series_1_robject)})

In [7]: df.rx2("series_1")
Out[7]:
R object with classes: ('AsIs',) mapped to:
<StrVector - Python:0x113a398c0 / R:0x7f8d13a8aec8>
[str, str, str, str]