.. _captureoutputrst:

=================================
Capture standard output and error
=================================


.. only:: html

    **Links:** :download:`notebook <capture_output.ipynb>`, :downloadlink:`html <capture_output2html.html>`, :download:`PDF <capture_output.pdf>`, :download:`python <capture_output.py>`, :downloadlink:`slides <capture_output.slides.html>`, :githublink:`GitHub|_doc/notebooks/capture_output.ipynb|*`


Python standard output is different from C standard output. When Python
embeds a C library which prints something on the standard output, it is
difficult to catch it from Python.

See also `print and
cleaning <https://docs.python.org/3/c-api/exceptions.html#printing-and-clearing>`__
and
`PySys_WriteStdout <https://docs.python.org/3/c-api/sys.html#PySys_WriteStdout>`__.

.. code:: ipython3

    from jyquickhelper import add_notebook_menu
    add_notebook_menu()






.. contents::
    :local:





.. code:: ipython3

    from cpyquickhelper.io import capture_output

Python Capture with Python print
--------------------------------

.. code:: ipython3

    def python_print():
        print("one line")
        print("two lines")
        
    res, out, err = capture_output(python_print, lang="py")
    type(out), out




.. parsed-literal::
    (str, 'one line\ntwo lines\n')



C Capture with C print
----------------------

.. code:: ipython3

    from cpyquickhelper.io.stdchelper import cprint
    
    def c_print():
        cprint("one line")
        cprint("two lines")
        
    res, out, err = capture_output(c_print, lang="c")
    type(out), out




.. parsed-literal::
    (bytes,
     b'o\x00n\x00e\x00 \x00l\x00i\x00n\x00e\x00t\x00w\x00o\x00 \x00l\x00i\x00n\x00e\x00s\x00')



Python capture with C print
---------------------------

.. code:: ipython3

    res, out, err = capture_output(c_print, lang="py")
    type(out), out




.. parsed-literal::
    (str, '')



C capture with Python print
---------------------------

On Windows, the behavior of this code is different in a standalone
program probably because jupyter catches the output on his side too.

.. code:: ipython3

    res, out, err = capture_output(python_print, lang="c")
    type(out), out


.. parsed-literal::
    one line
    two lines




.. parsed-literal::

    (NoneType, None)