cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Check out the JMP® Marketplace featured Capability Explorer add-in
Choose Language Hide Translation Bar
bjbreitling
bjbreitling
Level IV

Code to write JMP table to python

Created: Aug 6, 2020 05:35 PM  |  Last Modified: Jun 9, 2023 4:34 PM (4045 views)

bjbreitling_0-1596749134723.png

The above is some code I got from JMP documentation but when I run it nothing happens. I have 64-bit version of python 3.8 and JMP Pro 15. Can anyone help me with this?

 

Do I need to add numpy and scipy and pandas? I have those in my python downloads. Do I need to put them in a different folder too for JMP to query?

0 Kudos
11 REPLIES 11
Paul_Nelson
Staff Paul_Nelson
Staff

Re: Code to write JMP table to python

Apr 22, 2022 10:43 AM (529 views)  |  Posted in reply to message from bjbreitling 08-12-2020

64-bit JMP requires 64-bit Python 3, and 32-bit  JMP requires 32-bit Python 3.  Further any of the modules you install need to be installed for that version of Python, so if you had Numpy installed in the 32-bit version of Python, that version will NOT work with 64-bit JMP or 64-bit Python.

  

The issue is that JMP loads the Python shared library (.dll, .so, .dylib,...) into its process space.  As a result, Python and JMP both must be 32-bit or both 64-bit.  Numpy is one of the Python modules that heavily depends on native compiled code for the extension, rather than pure Python code which is system agnostic.  So it too must match the 32/64 bit-ness of JMP and Python.  Install Numpy, spicy, and pandas using the version of pip or pip3 located in the same directories as the Python you are using.

0 Kudos
Paul_Nelson
Staff Paul_Nelson
Staff

Re: Code to write JMP table to python

Jan 5, 2022 01:00 PM (542 views)  |  Posted in reply to message from bjbreitling 08-06-2020

JMP uses the pandas package, specifically a pandas data frame,  to send data tables back and forth with Python.  You need the missing Python packages for all of JMP's functionality to work properly.  For best success you should be using a Python.org version of Python.  With more than one version of Python on your system you will likely need to use the optional argument to the Python Init().  The bit-ness of the Python has to match the bit-ness of JMP.  64-bit JMP requires 64-bit Python, as JMP is loading the Python dll.  Anaconda is a lot more troublesome get to a working configuration, as it is designed to be the keeper, manager of everything Python.  It does not play nearly as nicely if you are trying to embed Python.

 

The additional packages needed by the first versions of JMP with Python support are numpy, pandas, scipy, matplotlib, and Qt5.  This list has been pared  down so that the most recent versions of JMP only require numpy, pandas, and matplotlib.  From your log warnings, you need numpy, scipy and pandas.

 

 Use Python's package manager pip ( or pip3 ). Make sure you use the pip in the same directory tree as the Python version you want to use with JMP.

 

pip install numpy scipy pandas

 

This will download and install the Python packages from public Python repositories and install them on your machine.

0 Kudos