pandas and rpy2: Why does ezANOVA work via robjects.r but not robjects.packages.importr?

Question

Like many, I'm hoping to stop straddling R and Python worlds and just work in Python using Pandas, Pyr2, Numpy, etc. I'm using the R package ez for its ezANOVA facility. It works if I do things the hard way, but why doesn't it work when I do them the easy way? I don't understand the resulting error:

File "/Users/malcomreynolds/analysis/r_with_pandas.py", line 38, in <module>
    res = ez.ezANOVA(data=testData, dv='score', wid='subjectid', between='block', detailed=True)
  File "/usr/local/lib/python2.7/site-packages/rpy2/robjects/functions.py", line 178, in __call__
    return super(SignatureTranslatedFunction, self).__call__(*args, **kwargs)
  File "/usr/local/lib/python2.7/site-packages/rpy2/robjects/functions.py", line 106, in __call__
    res = super(Function, self).__call__(*new_args, **new_kwargs)
rpy2.rinterface.RRuntimeError: Error in table(temp[, names(temp) == wid]) : 
  attempt to set an attribute on NULL

See below for full reproducible code (requires some python packages: pyr2, pandas, numpy):

import pandas as pd
from rpy2 import robjects
from rpy2.robjects import pandas2ri
pandas2ri.activate()  # make pyr2 accept and auto-convert pandas dataframes
from rpy2.robjects.packages import importr
base = importr('base')
ez = importr('ez')
robjects.r['options'](warn=-1)  # ???
import numpy as np

"""Make pandas data from from scratch"""

score = np.random.normal(loc=10, scale=20, size=10)
subjectid = range(10)
block = ["Sugar"] * 5 + ["Salt"] * 5
testData = pd.DataFrame({'score':score, 'block':block, 'subjectid': subjectid})
# it looks just like a dataframe from R
print testData

"""HARD WAY: Use ezANOVA thorugh pyr2 *** THIS WORKS ***"""
anova1 = robjects.r("""
library(ez)
function(df) {
    # df gets passed in
    ezANOVA(
        data=df,
        dv=score,
        wid=subjectid,
        between=block,
        detailed=TRUE)
}
""")
print anova1(testData)


# this command shows that ez instance is setup properly
print ez.ezPrecis(data=testData)  # successful

"""EASY WAY: Import ez directly and use it """
# *** THIS APPROACH DOES NOT WORK ***
# yet, trying to use ez.ezANOVA yields an excpetion aboutthe wid value
# res = ez.ezANOVA(data=testData, dv='score', wid='subjectid', between='block', detailed=True)
# print res

# *** THIS APPROACH WORKS (and also uses my options change) ***
res = ez.ezANOVA(data=testData, dv=base.as_symbol('score'), wid=base.as_symbol('subjectid'), between=base.as_symbol('block'))
print res

Answer 1

In the easy version you are passing symbol names as strings. This is not the same as a symbol.

Check the use of as_symbol in Minimal example of rpy2 regression using pandas data frame