apache / apache/arrow

[Python] Python install pyarrow failed on Linux ppc64le

Open
#18,997 0 comments 0 reactions 0 assignees View on GitHub
Component: Python installation Type: bug
Dominant language
C++
Stars
17.1k
Forks
4.3k
Avg merge
3d 13h
Merged PRs (30d)
88

Description

I found a similar problem issue from [StackOverflow](https://stackoverflow.com/questions/66902732/pip-install-pyarrow-failed-on-linux-ppc64le)

I am using Python 3.7 and PIP 21.0 and linux ppc64le in a cloud instance-based. I tried to install PyArrow with conda-forge but has the similar errors stated

 

**Incompatible with system CUDA Driver 11.0.0**

**ERROR: Failed building wheel for pyarrow Failed to build pyarrow ERROR: Could not build wheels for pyarrow which use PEP 517 and cannot be installed directly**

 

I want to know does PyArrow supports linux ppc64le? Even the conda page said it does, but I tried to install different versions of PyArrow but keep getting the same error. Does that mean I need to reinstall or downgrade my CUDA driver?

 

Any advice will be much appreciated!

**Environment**: linux ppc64le in the cloud-based
**Reporter**: [Justin](https://issues.apache.org/jira/browse/ARROW-15120)

**Note**: *This issue was originally created as [ARROW-15120](https://issues.apache.org/jira/browse/ARROW-15120). Please see the [migration documentation](https://github.com/apache/arrow/issues/14542) for further details.*

Contributor guide

Open the contributing guide

Research direction

No project file or test is identified. Start by reproducing the installation with Python 3.7, pip 21.0, Linux ppc64le, and the reported CUDA 11.0 driver, comparing pip and conda-forge results. Done means determining whether PyArrow supports this environment and documenting the required compatible versions or installation path.

Written by the indexing model from the issue text.

Assessment

Tech stack
linux, python
Domain
build-system
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.