topic Re: How to write .jmp files with JMP python in Discussions

How to write .jmp files with JMP python

AdrianBWeissman — Tue, 02 Jun 2026 17:31:59 GMT

Hi, I am using JMP19.1.1

I see examples of how to use JMP or JSL to open files, add jsl scripts for plotting and then run them.

Are there examples to use JMP Python to open csv files and write .jmp files with charts in them? What do people do?

Adrian

Re: How to write .jmp files with JMP python

jthi — Tue, 02 Jun 2026 18:34:38 GMT

Saving can be done with save method of DataTable

To add table scripts you can use either run_jsl or scripts with jmp.Expression

Re: How to write .jmp files with JMP python

AdrianBWeissman — Tue, 02 Jun 2026 19:50:32 GMT

Hi,

Thanks for responding. I am talking about not using the jmp libraries. I am talking about using python code that is processing data, that spits out csv files. I want to switch to writing .jmp files. Is there a way to do this? i.e. Pandas dataframe directly to .jmp ?

Thanks

Re: How to write .jmp files with JMP python

jthi — Wed, 03 Jun 2026 04:26:39 GMT

To my knowledge no. You would need to run the Python inside JMP.

Re: How to write .jmp files with JMP python

AdrianBWeissman — Wed, 03 Jun 2026 14:26:45 GMT

Here is what I try to run:

from pathlib import Path
import jmp

csv_path = Path(r"C:\Work\debug\jmp_debug\CT_Pro_PRO0CV2_pass_fail_05302026.csv")
out_path = Path(r"C:\Work\debug\jmp_debug\CT_Pro_PRO0CV2_pass_fail_05302026.jmp")

print("Opening:", csv_path)

dt = jmp.open(str(csv_path))

print("Saving:", out_path)

dt.save(str(out_path))

print("Done")

I run it this way: & "C:\Program Files\JMP\JMP\19\jmp.exe" -R "C:\Work\debug\jmp_debug\jmp_csv_to_jmp_simple.py" in powershell. The problem is this pops up three windows. Do you know if there is anyway to run JMP in a headless fashion? I.e. Run it and not have a collage of windows pop open ?

Re: How to write .jmp files with JMP python

AdrianBWeissman — Wed, 03 Jun 2026 14:31:44 GMT

The main reason to go directly to .jmp is these files are 4-10 GB in csv format. So this is painful. Ideally, we could pass a pandas dataframe to the jmp table and write it

Re: How to write .jmp files with JMP python

AdrianBWeissman — Wed, 03 Jun 2026 14:44:04 GMT

This is really painful. It appears there is no real way to make this run in a headless fashion. This is really surprising. I must me wrong and not have the right documenation. If someone has information on how to run jmp in a headless fashion, this would be really helpful. I.e. I can run this in a script and not get three to four new windows popping open.

Re: How to write .jmp files with JMP python

AdrianBWeissman — Wed, 03 Jun 2026 15:17:41 GMT

In JMP 19.1.3, and plain vanilla Python 3.13.3 (Non JMP), I got the following improvement. This only pops up the .jmp table being created. So this is close to a solution. I would really still like to know if there is headless information ?

from pathlib import Path
import time

import win32com.client

def jsl_quote(path_or_text) -> str:
text = str(path_or_text)
text = text.replace("\\", "\\\\").replace('"', '\\"')
return f'"{text}"'

csv_path = Path(r"C:\Work\debug\jmp_debug\05302026.csv")
out_path = Path(r"C:\Work\debug\jmp_debug\05302026.jmp")

jmp = win32com.client.Dispatch("JMP.Application")

# Try False first. If nothing works, set True while debugging.
jmp.Visible = False

jsl = f"""
Names Default To Here( 1 );

dt = Open({jsl_quote(csv_path)},Invisible);

dt << Save As( {jsl_quote(out_path)} );

Close( dt, No Save );
"""

print("Running JSL...")
jmp.RunCommand(jsl)

# Give JMP a moment if RunCommand returns before save completes.
time.sleep(2)

print("Saved?", out_path.exists(), out_path)

Re: How to write .jmp files with JMP python

Paul_Nelson — Wed, 03 Jun 2026 20:35:46 GMT

CSV for large files is painful. My suggestion is to write out the files in parquet format, then import pyarrow in JMP, load the parquet file, it can be converted to a pandas dataframe with almost no time cost, JMP can directly read pandas columns in building a data table and has a jmp.from_dataframe( df ) capability. If you are using pandas 3.0 or greater they have started some deprecation warnings, but those have been take care of in development versions of JMP by directly using the pyarrow PyCapsule interface. CSV also has the issue of not being an exact format. Check your decimal points. Conversion from string to double can have small differences especially since it's from numeric to CSV and then back again.