Failed to install pandas==2.3.0 with Python 3.9 #61563
Comments
ryanchao2012
added
Build
|
Ran into the same issue in Mac OSX (Sillicon), using uv pip install. Error is in Cython there is a missing "ios" file. Python: 3.12.7 Most relevant log: |
|
See #61564 (comment) The new version of pandas (2.3.0) requires python version >=3.10. I ran into the same issue this morning in my Gitlab Pipeline Runners that were using python 3.9-slim image and failing. I fixed the pandas version to 2.2.3 in my requirements.txt file and it worked. Could alternatively (and should) upgrade my runners to run using Python 3.10+ |
|
I've faced a similar issue using Relevant bit from the generated dockerfile. RUN apt install -y python3.9 python3.9-distutils \
&& ln -s -f $(which python3.9) /usr/bin/python \
&& wget https://bootstrap.pypa.io/get-pip.py -O /tmp/get-pip.py \
&& python /tmp/get-pip.py
RUN pip install mlflow==2.22.0I suggest to stop using |
|
I am seeing the same issue.
as you can see, it's missing cpython releases for 3.11, 3.12. |
|
We are also seeing the issue; the build log shows: |
mroeschke
changed the title
|
Thanks for the report. For those asking about potentially missing wheels for Python >=3.10 on PyPI, I just uploaded a few just now #61564 (comment) Regarding 3.9, pandas 2.3.0 intended to drop support for this version (#60792) however the |
|
cc @pandas-dev/pandas-core if there are any considerations I'm missing above |
You'd have to update the I believe that there may be things in the source that depend on 3.10. Did the 3.9 wheels pass our tests? |
Yup appears the wheels are passing the test suite e.g. https://github.com/pandas-dev/pandas/actions/runs/15473507739/job/43563581403?pr=61569 |
Maybe I'm missing something, but I don't see where |
|
so is it supported or not supported in python 3.9 ? |
You'll need to expand the Specifically here is where pytest is run against OSX 3.9 wheels in the logs: https://github.com/pandas-dev/pandas/actions/runs/15473507739/job/43563581403?pr=61569#step:7:10168 (The tests are a subset of the full test suite e.g. not the slow, network or db tests) Since the wheel tests are passing, I'm leaning toward just uploading the 3.9 wheels to PyPI, especially since 3.0 will be the next pandas version. |
|
We are running into this issue as well in the Plotly CI; since +1 to uploading 3.9 wheels; otherwise we will need to pin |
Thanks for the pointer. I see it now.
OK - the docs on the web site have this (but it is not referring to the version): https://pandas.pydata.org/docs/whatsnew/v2.3.0.html#increased-minimum-version-for-python So I think they need a fresh upload. |
I think that is indeed the best option, because even if we would do a new 2.3.1 release that fixes the metadata, I think pip will still find 2.3.0 as the latest release supporting Py 3.9, so people would otherwise keep running into this. |
|
Personally, I'd probably update |
|
OK I'll go ahead and just upload the 3.9 wheels to PyPI using the wheels from this PR #61569 (I don't feel great doing a 2.3.1 release with the bumpiness of this release so far). |
|
The Python 3.9 wheels have been uploaded to PyPI: https://pypi.org/project/pandas/2.3.0/#files |
With those files now available, should we expect users of pandas 2.3.0 / python 3.9 to "just work"? That is, will package management systems automagically obtain the correct corresponding wheel for pandas 2.3.0 (given the python version, that is)? |
Yes, it works. My failing python 3.9 pipeline works as intended now. |
|
Hey @mroeschke , thanks for the quick fix, it's working now. |
|
Working on our side. Thanks for the quick fix @mroeschke 🎉 |
Installation check
Platform
Linux-5.10.195-1.20230921.el7.x86_64-x86_64-with-glibc2.17
Installation Method
pip install
pandas Version
2.3.0
Python Version
3.9.15
Installation Logs
(base) [root@64bf929a621d7dafeb18b348 ~]# pip install pandas -U
Requirement already satisfied: pandas in /opt/conda/lib/python3.9/site-packages (1.5.0)
Collecting pandas
Downloading pandas-2.3.0.tar.gz (4.5 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 4.5/4.5 MB 90.9 MB/s eta 0:00:00
Installing build dependencies ... done
Getting requirements to build wheel ... done
Installing backend dependencies ... done
Preparing metadata (pyproject.toml) ... error
error: subprocess-exited-with-error
× Preparing metadata (pyproject.toml) did not run successfully.
│ exit code: 1
╰─> [152 lines of output]
+ meson setup /tmp/pip-install-gp_gpioe/pandas_8608342ddb164d0e8725d2463640de86 /tmp/pip-install-gp_gpioe/pandas_8608342ddb164d0e8725d2463640de86/.mesonpy-lsu89q1a -Dbuildtype=release -Db_ndebug=if-release -Db_vscrt=md --vsenv --native-file=/tmp/pip-install-gp_gpioe/pandas_8608342ddb164d0e8725d2463640de86/.mesonpy-lsu89q1a/meson-python-native-file.ini
The Meson build system
Version: 1.8.1
Source dir: /tmp/pip-install-gp_gpioe/pandas_8608342ddb164d0e8725d2463640de86
Build dir: /tmp/pip-install-gp_gpioe/pandas_8608342ddb164d0e8725d2463640de86/.mesonpy-lsu89q1a
Build type: native build
Project name: pandas
Project version: 2.3.0
C compiler for the host machine: cc (gcc 4.8.5 "cc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)")
C linker for the host machine: cc ld.bfd 2.27-44
C++ compiler for the host machine: c++ (gcc 4.8.5 "c++ (GCC) 4.8.5 20150623 (Red Hat 4.8.5-44)")
C++ linker for the host machine: c++ ld.bfd 2.27-44
Cython compiler for the host machine: cython (cython 3.1.1)
Host machine cpu family: x86_64
Host machine cpu: x86_64
Program python found: YES (/opt/conda/bin/python)
Found pkg-config: YES (/usr/bin/pkg-config) 0.27.1
Run-time dependency python found: YES 3.9
Build targets in project: 53
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
The text was updated successfully, but these errors were encountered: