DLL load failed while importing turbodbc_intern

[swolfe2]

Our systems team have forced an upgrade from Python 3.7 to the newest 3.9.6. I was able to successfully install turbodbc with Boost libraries. However, when I try to run an existing script that worked on Python 3.7, I get the following error:

C:\Users\UserID\Desktop\Code>"C:/Program Files/Python39/python.exe" "c:/Users/UserID/Desktop/Code/Python/Warehouse IDs.py"
Traceback (most recent call last):
  File "c:\Users\UserID\Desktop\Code\Python\Warehouse IDs.py", line 31, in <module>
    from turbodbc import connect, make_options
  File "C:\Users\UserID\AppData\Roaming\Python\Python39\site-packages\turbodbc\__init__.py", line 2, in <module>
    from .connect import connect
  File "C:\Users\UserID\AppData\Roaming\Python\Python39\site-packages\turbodbc\connect.py", line 1, in <module>
    from turbodbc_intern import connect as intern_connect
ImportError: DLL load failed while importing turbodbc_intern: A dynamic link library (DLL) initialization routine failed.

Is there something that I'm missing to get it to work with the newest version of Python?

[dhirschfeld]

You have missing or incompatible dependencies. IMO the proper solution to fix DLL load issues on Windows is to use mamba and conda-forge - that's exactly what they were designed to do. pip was never designed to install packages with complex compiled dependencies.

If you can't use conda then you may or may not be able to get information about what the problem is by inspecting the pyd file with dependencies.

[xhochy]

This is a duplicate of #258, would be interesting to find out the cause.

In addition to dependencies, one could also check out the newly released https://github.com/microsoft/dlltracer-python to see where this breaks.

[swolfe2]

Hello @xhochy, apologies for the delay in response. I had to submit a request to get dlltracer to work with newer versions of Python so I could also test it. Below is a an output of the dlltracer run for turbodbc on the newest Python 3.9.6 I have on my system:

It would be awesome if I could get turbodbc to work with it, as it is by far the fastest method I've used to do work in MSSQL with Python.

Appreciate it!