ImportError: DLL load failed

[fluviotect]

I got an import error running your second example in Jupyter Lab (Py 3.7 Win7x64).

I crossed-referenced packages in my Conda env with dependencies listed on your GitHub page and two were missing:

• cython
• scikit-build

So it seems when I installed pdal, some but not all of the dependencies missing from my env were identified and/or installed.

If it matters, to install pdal, I used:

conda install -c conda-forge python-pdal

Attempting to resolve, I installed the two missing packages (and their dependencies). However, I still get the same error when I run notebook.

---

ImportError Traceback (most recent call last)

in 12 }""" 13 ---> 14 import pdal 15 import numpy as np 16 pipeline = pdal.Pipeline(json % data) C:\ProgramData\Anaconda3\envs\py37\lib\site-packages\pdal\__init__.py in 1 __version__='2.3.6' 2 ----> 3 from .pipeline import Pipeline 4 from .array import Array 5 from .dimension import dimensions C:\ProgramData\Anaconda3\envs\py37\lib\site-packages\pdal\pipeline.py in 1 ----> 2 from pdal import libpdalpython 3 import numpy as np 4 5 class Pipeline(object): ImportError: DLL load failed: The specified procedure could not be found.

[hobu]

I crossed-referenced packages in my Conda env with dependencies listed on your GitHub page and two were missing:

• cython
• scikit-build

Were these needed to run the PDAL Python bindings or just to build? They are requirements for building, but they shouldn't be for running.

When I create a fresh environment, python -m pdal works fine for me on Win10:

conda create -n issue-85
conda activate issue-85
conda install python=3.7 python-pdal

(issue-85) C:\Users\hobu\Desktop>python -m pdal
C:\Miniconda3\envs\issue-85\python.exe: No module named pdal.__main__; 'pdal' is a package and cannot be directly executed

It seems your environment is resolving an older python-pdal package. You should try to upgrade just that.

python-pdal               2.4.0            py37h0b66150_2    conda-forge