[Python] Problem importing PyArrow on a cluster

[asfimport]

Hi, when I am trying to import pyarrow in python, I get the following error:

File "", line 1, in
File "/services/matejm/anaconda3/lib/python3.7/site-packages/pyarrow/init.py", line 49, in
from pyarrow.lib import cpu_count, set_cpu_count
ImportError: libaws-cpp-sdk-s3.so: cannot open shared object file: No such file or directory
What can this be related to? I have searched wherever i could've and could not find any reason for it, so I figured i might as well try in here.

Thank you very much

Note: I have installed py-arrow and its dependencies off-line, since our cluster has a company firewall that does not allow pip nor conda installation

Environment: Linux, RedHat CentOS 7
Reporter: Matej Murin

_{Note: This issue was originally created as ARROW-8135. Please see the migration documentation for further details.}

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
Can you give more details on how you install pyarrow and its dependencies?

[asfimport]

Matej Murin:
I have done:

**conda info pyarrow**

Locally on my machine.
Found my version, and checked if I have all the dependencies listed there in my conda list command.
The ones i didn't, I found the dependencies on conda-forge website, downloaded the tar.bz2 files manually and transferred them, then Installed offline.

[asfimport]

Wes McKinney / @wesm:
It looks like you're missing transitive dependencies of pyarrow. You also need the runtime dependencies of arrow-cpp, the C++ library

https://github.com/conda-forge/arrow-cpp-feedstock/blob/master/recipe/meta.yaml#L50

Recommend closing the issue

[asfimport]

Matej Murin:
@wesm yes, thank you for the link 

[asfimport]

Matej Murin:
Had been missing dependencies of the package