rpy2 / R interface (v0.16.0)

[borisRa]

Hi,

Using the last pandas version, the command : "import pandas.rpy.common as com"

produces the following warning :
"The pandas.rpy module is deprecated and will be removed in a future version. We refer to external packages like rpy2, found here: http://rpy.sourceforge.net "like rpy2, found here: ttp://rpy.sourceforge.net", FutureWarning) "

I use the following functionality heavily :
com.convert_to_r_dataframe(pandas_DF)
com.convert_robj(results_from_R)

How with rpy2 I can do the same ?

Regards,
Boris

[jorisvandenbossche]

See #9602.
We were actually a little bit too fast with the deprecation, before we could put good docs in place on how to port to rpy2 (and the docs in rpy2 itself on this are also not very clear)

I think you can start with this: http://stackoverflow.com/questions/20630121/pandas-how-to-convert-r-dataframe-back-to-pandas
And we would really appreciate some feedback/problems you encounter (eg things that work in pandas but not with rpy2), you can report that at #9602

[jreback]

closing in favor of #9602

[jorisvandenbossche]

@borisRa Here is a notebook exploring some things: http://nbviewer.ipython.org/gist/jorisvandenbossche/d0deaa53ace697e1514e

Basically:

• imports:

  import rpy2.robjects.conversion as conversion
  from rpy2.robjects import pandas2ri
  pandas2ri.activate()
  

• after that, you can change:

• com.convert_to_r_dataframe(df) -> conversion.py2ri(df)

• com.convert_robj(rdf) -> conversion.ri2py(rdf)

Does this work for you?

[borisRa]

Thanks ! that is exactly what I looked for !

[borisRa]

And what are the alternatives for :

import pandas.rpy.common as com
pydf = com.load_data('iris')
r_matrix = com.convert_to_r_matrix(df)

Thanks,
Boris

[jorisvandenbossche]

For the data, does this work?

from rpy2.robjects import r
r.data(name)
robj = r[name]
conversion.ri2py(robj)

[jorisvandenbossche]

And for the matrix, this is actually just using convert_to_r_dataframe, and then using the R as.matrix function.

[borisRa]

Everything works as a charm !
Thanks !