Use Python with R with reticulate :: Cheatsheet

Hex logo for reticulate - a navy blue background with a light-blue and yellow snake semi-coiled across the foreground. 'reticulate' is written across the bottom in white.

Download PDF

Translations (PDF)

The reticulate package lets you use Python and R together seamlessly in R code, in R Markdown documents, and in the RStudio IDE.


Python in R Markdown

(Optional) Build Python env to use.

knitr versions >= 1.18 will automatically use the reticulate engine for Python chunks. See ?reticulate::eng_python for a listing of supported knitr chunk options.

#| label: setup
#| include: false

py_install("seaborn", envname = "fmri-proj") 

Begin Python chunks with ```{python}. Chunk options like echo, include, etc. all work as expected.

#| echo: false

import seaborn as sns 
fmri = sns.load_dataset("fmri")`

Use the py object the access objects created in Python chunks from R chunks.

``` {{r}}
f1 <- subset(py$fmri, region = "parietal")

Python chunks all execute within a single Python session so you have access to all objects created in previous chunks.

Use the r object to access objects created in R chunks from Python chunks.

``` {{python}}
import matplotlib as mpl
sns.lmplot("timepoint", "signal", data=r.f1)

Python in R

python.r Example:


sns <- import("seaborn")

fmri <- sns$load_dataset("fmri")

# creates tips

# creates tips in main

py_run_string("print(tips.shape)") Example:

import seaborn as sns
tips = sns.load_dataset("tips")

Call Python from R code in three ways:

Import Python Modules

Use import() to import any Python module. Access the attributes of a module with $.


sns <- import("seaborn")

tips <- sns$load_dataset("tips")
  • import(module, as = NULL, convert = TRUE, delay_load = FALSE): Import a Python module. If convert = TRUE, Python objects are converted to their equivalent R types. Also import_from_path().

  • import_main(convert = TRUE): Import the main module, where Python executes code by default.

  • import_builtings(convert = TRUE): Import Python’s built-in functions.


Source Python Files

Use source_python() to source a Python script and make the Python functions and objects it creates available in the calling R environment.

  • source_python(file, envir = parent.frame(), convert = TRUE): Run a Python script, assigning objects to a specified R environment.


Run Python Code

Execute Python code into the main Python modules with py_run_file() or py_run_string().


  • py_run_string(code, local = FALSE, convert = TRUE): Run Python code (passed as a string) in the main module.

    py_run_string("x = 10")
  • py_run_file(file, local = FALSE, convert = TRUE): Run Python file in the main module.

  • py_eval(code, convert = TRUE): Run a Python expression, return the result. Also py_call().

    py_eval("1 + 1")

Access the results, and anything else in Python’s main module, with py.

  • py: An R object that contains the Python main module and the results stored there.


Object Conversion

Tip: To index Python objects begin at 0, use integers, e.g. OL

Reticulate provides automatic built-in conversion between Python and R for many Python types.

Table of data types in R and their Python equivalents.
R Python
Single-element vector Scalar
Multi-element vector List
List of multiple types Tuple
Named List Dict
Matrix/Array NumPy ndarray
Data Frame Pandas DataFrame
Function Python function
NULL, TRUE, FALSE None, True, False

Or, if you like, you can convert manually with

  • py_to_r(x): Convert a Python object to an R object. Also r_to_py().

  • tuple(..., convert = FALSE): Create a Python tuple.

    tuple("a", "b", "c")
  • dict(..., convert = FALSE): Create a Python dictionary object. Also py_dict() to make a dictionary that uses Python objects as keys.

    dict(foo = "bar", index = 42L)
  • np_array(data, dtype = NULL, order = "C"): Create NumPy arrays.

    np_array(c(1:8), dtype = "float16")
  • array_reshape(x, dim, order = c("C", "F")): Reshape a Python array.

    x <- 1:4
    array_reshape(x, c(2,2))
  • py_func(f): Wrap an R function in a Python function with the same signature.

  • py_main_thread_func(f): Create a function that will always be called on the main thread.

  • iterate(it, f = base::identity, simplify = TRUE): Apply an R function to each value of a Python iterator or return the values as an R vector, draining the iterator as you go. Also iter_next() and as_iterator().

    iterate(iter, print)
  • py_interator(fn, completed = NULL): Create a Python iterator from an R function.

    seq_gen <- function(x) {
      n <- x;
      function() {
        n <<- n + 1;


  • py_capture_output(expr, type = c("stdout", "stderr")): Capture and return Python output. Also py_supress_warnings().

  • py_get_attr(x, name, silent = FALSE): Get an attribute of a Python object. Also py_set_attr(), py_has_attr(), and py_list_attributes().

  • py_help(object): Open the documentation page for a Python object.

  • py_last_error(): Get the last Python error encountered. Also py_clear_last_error() to clear the last error.

  • py_save_object(object, filename, pickle = "pickle", ...): Save and load Python objects with pickle. Also py_load_object().

    py_save_objects(x, "x.pickle")
  • with(data, expr, as = NULL, ...): Evaluate an expression within a Python context manager.

    py <- import_builtins()
    with(py$open("output.txt", "w") %as% file,
         {file$write("Hello, there!")})

Python in the IDE

Requires reticulate plus RStudio v1.2+. Some features require v1.4+.

reticulate features in the RStudio IDE

  • Syntax highlighting for Python scripts and chunks.
  • Tab completion for Python functions and objects (and Python modules imported in R scripts).
  • Source Python scripts.
  • Execute Python code line by line with Cmd + Enter (Ctrl + Enter).
  • View Python objects in the Environment Pane.
  • View Python objects in the Data Viewer.
  • A Python REPL opens in the console when you run Python code with a keyboard shortcut. Type exit to close.
  • matplotlib plots display in plots pane.
  • Press F1 over a Python symbol to display the help topic for that symbol.

Python REPL

RStudio IDE Window:

A REPL (Read, Eval, Print Loop) is a command line where you can run Python code and view the results.

  1. Open in the console with repl_python(), or by running code in a Python script with Cmd + Enter (Ctrl + Enter).

    • repl_python(module = NULL, quiet = getOption("reticulate.repl.quiet", default = FALSE), input = NULL): Launch a Python REPL. Run exit to close.

  2. Type commands at >>> prompt.

  3. Press Enter to run code.

  4. Type exit to close and return to R console.

    > reticulate::repl_python()
    Python 3.9.16 (/Users/mine/.virtualenvs/r-reticulate/bin/python)
    Reticulate 1.28 REPL -- A Python interpreter in R.
    Enter 'exit' or 'quit' to exit the REPL and return to R.
    >>> import seaborn as sns
    >>> tips = sns.load_dataset("tips")
    >>> tips.shape
    (244, 7)
    >>> exit

Configure Python

Reticulate binds to a local instance of Python when you first call import() directly or implicitly from an R session. To control the process, find or build your desired Python instance. Then suggest your instance to reticulate. Restart R to unbind.

Find Python

  • install_python(version, list = FALSE, force = FALSE): Download and install Python.

  • py_available(initialize = FALSE): Check if Python is available on your system. Also py_module_available() and py_numpy_module().

  • py_discover_config(): Return the detected installation of Python. Use py_config() to check which version has been loaded.

  • virtualenv_list(): List all available virtual environments. Also virtualenv_root().

  • conda_list(conda = "auto"): List all available conda envs. Also conda_binary() and conda_version().


Create a Python env

  • virtualenv_create(envname = NULL, ...): Create a new virtual environment.

  • conda_create(envname = NULL, ...): Create a new conda environment.

    conda_create("r-pandas", packages = "pandas")

Install Packages

Install Python packages with R (below) or the shell:

pip install SciPy

conda install SciPy

  • py_install(packages, envname, ...): Install Python packages into a Python env.

  • virtualenv_install(envname, packages, ...): Install a package within a virtual environment. Also virtualenv_remove().

    virtualenv_install("r-pandas", packages = "pandas")
  • conda_installs(envname, packages, ...): Install a package within a conda environment. Also conda_remove().

    conda_install("r-pandas", packages = "plotly")

Suggest an env to use

Set a default Python interpreter in the RStudio IDE Global or Project Options. Go to Tools > Global Options … > Python for Global Options. Within a project, go to Tools > Project Options… > Python.

Otherwise, to choose an instance of Python to bind to, reticulate scans the instances on your computer in the following order, stopping at the first instance that contains the module called by import().

  1. The instance referenced by the environment variable RETICULATE_PYTHON (if specified). Tip: set in .Renviron file.

    • Sys.setenv(RETICULATE_PYTHON = PATH): Set default Python binary. Persists across sessions! Undo with Sys.unsetenv().

      Sys.setenv(RETICULATE_PYTHON = "/usr/local/bin/python")
  2. The instances referenced by use_ functions if called before import().

    • use_python(python): Path to a Python binary.

    • use_virtualenv(virtualenv): Path to or name of a Python virtualenv.

      #| eval: false
      use_virtualenv("~/myenv") # path to venv
      use_virtualenv("r-keras") # name of venv
  3. A virtual env found in the current working directory: “./.venv”

  4. Environments that are named after the imported module. e.g. “~/.virtualenvs/r-scipy/” for import("scipy")

  5. The package default virtualenv, “r-reticulate”.

  6. At the location of the Python binary discovered on the system PATH (i.e. Sys.which("python"))

CC BY SA Posit Software, PBC •

Learn more at

Updated: 2024-06.

[1] '1.37.0'