Rstudio proper settings for work with Python?

[marthinek]

I installed standalone Rstudio version with R.

Separately to this, I have Python environment setup with anaconda.
Python environment works with no issues in jupyter notebook, jupyter lab, spyder IDE or under VScode.

Now I want to use R-markdown for reporting.
In R studio, I start New Project.

in Tools Menu, Project Options and Global Options, under Python section I choose my conda environment

then I create new R-markdown file in my project directory.

I can knit the template with no issues.

Now I add python chunks:

1. environment check:

# check environment
Sys.getenv("RETICULATE_PYTHON")

it returns: [1] "C:/Users/.../.conda/envs/env_expanalysis/python.exe"

also python chunk:

x=42
print(x)

runs with no issues (with both 'run current chunk' and knit).

So, to me it looks like python environment is setup correctly.

I can even plot something with matplotlib running this chunk

import matplotlib.pyplot as plt
import numpy as np

xpoints = np.array([1, 8])
ypoints = np.array([3, 10])

plt.plot(xpoints, ypoints)
plt.show()

This works fine.

Now, I need pandas library for my analysis.. and here is the problem

import pandas as pd

this returns an error:
"ImportError: DLL load failed while importing _bz2: The specified module could not be found."

Of course pandas is installed and works fine in jupyter notebook using the same environment.

Do you have any ideas what is wrong here?
Thanks.

[cderv]

@marthinek

this returns an error:
"ImportError: DLL load failed while importing _bz2: The specified module could not be found."

If I look for this, it seems to be an issue related with the python environment

official embedded Python fails to import certain modules

Are you using 3.8 in this environment ?
But you said it is working in jupyter notebook with the same python version right ?

knitr uses reticulate to run the python chunk, you could try running python from R with reticulate directly from R console (outside of R markdown) to check that it works.

It seems to me something cause pandas loading to fail in the Python installation.