topic Seaborn breaks Pandas in ArcGIS Pro Questions

Seaborn breaks Pandas

ChristopherBater — Fri, 26 Nov 2021 14:44:24 GMT

Hi, I'm using Pro 2.8.2 and Spyder 5 as my IDE. Adding Seaborn to a cloned environment breaks Pandas. When I try to import Pandas I get an error (see below); importing Seaborn causes a similar error. Any thoughts?

Thanks,

Chris

In [1]: import pandas
Traceback (most recent call last):

File "<ipython-input-1-38d4b0363d82>", line 1, in <module>
import pandas

File "C:\Users\user.name\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\__init__.py", line 22, in <module>
from pandas.compat import (

File "C:\Users\user.name\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\compat\__init__.py", line 23, in <module>
from pandas.compat.pyarrow import (

File "C:\Users\user.name\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\compat\pyarrow.py", line 9, in <module>
_palv = Version(_pa_version)

File "C:\Users\user.name\AppData\Local\ESRI\conda\envs\arcgispro-py3-clone1\lib\site-packages\pandas\util\version\__init__.py", line 339, in __init__
match = self._regex.search(version)

TypeError: expected string or bytes-like object

Re: Seaborn breaks Pandas

DanPatterson — Fri, 26 Nov 2021 17:01:00 GMT

What are the specific installed versions of matplotlib, pandas and pyarrow? and what version of seaborn are you trying to use? There may be a version conflict. Also, you didn't use pip anywhere during the install process did you?

Re: Seaborn breaks Pandas

DanPatterson — Fri, 26 Nov 2021 18:03:24 GMT

Without seaborn installed, here are the packages and versions

# Name Version Build Channel freetype 2.10.4 vc14_3 [vc14] esri (arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>conda list matplotlib # packages in environment at C:\arc_pro\bin\Python\envs\arcgispro-py3: # # Name Version Build Channel matplotlib 3.4.2 py37_arcgispro_1 [arcgispro] esri (arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>conda list pandas # packages in environment at C:\arc_pro\bin\Python\envs\arcgispro-py3: # # Name Version Build Channel pandas 1.3.4 py37h6214cd6_0 (arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>conda list spyder # packages in environment at C:\arc_pro\bin\Python\envs\arcgispro-py3: # # Name Version Build Channel pyls-spyder 0.4.0 pyhd3eb1b0_0 spyder 5.1.5 py37haa95532_1 spyder-kernels 2.1.3 py37haa95532_0 (arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>

and here is what it wants to install

(arcgispro-py3) C:\arc_pro\bin\Python\envs\arcgispro-py3>conda install seaborn --dry-run Collecting package metadata (current_repodata.json): done Solving environment: done ## Package Plan ## environment location: C:\arc_pro\bin\Python\envs\arcgispro-py3 added / updated specs: - seaborn The following packages will be downloaded: package | build ---------------------------|----------------- seaborn-0.11.2 | pyhd3eb1b0_0 218 KB ------------------------------------------------------------ Total: 218 KB The following NEW packages will be INSTALLED: seaborn pkgs/main/noarch::seaborn-0.11.2-pyhd3eb1b0_0 DryRunExit: Dry run. Exiting.

Re: Seaborn breaks Pandas

ChristopherBater — Wed, 01 Dec 2021 17:14:16 GMT

Hi Dan, I installed Seaborn using Pro's Python package manager (ie not using Conda' command prompt). As requested, see below, and thanks for your time!

Without Seaborn # Name Version Build Channel matplotlib 3.3.1 py37_arcgispro_0 [arcgispro] esri pandas 1.2.3 py37hf11a4ad_0 pyarrow 1.0.1 1 esri pyls-spyder 0.4.0 pyhd3eb1b0_0 spyder 5.1.5 py37haa95532_1 spyder-kernels 2.1.3 py37haa95532_0 With Seaborn # Name Version Build Channel matplotlib 3.3.1 py37_arcgispro_0 [arcgispro] esri pandas 1.3.4 py37h6214cd6_0 pyarrow 1.0.1 1 esri pyls-spyder 0.4.0 pyhd3eb1b0_0 seaborn 0.11.2 pyhd3eb1b0_0 spyder 5.1.5 py37haa95532_1 spyder-kernels 2.1.3 py37haa95532_0

Re: Seaborn breaks Pandas

DanPatterson — Wed, 01 Dec 2021 18:05:21 GMT

I have matplotlib 3.4.2 yours is 3.3.1, I haven't installed Seaborn in a while, but the version is different and you will only find out if it should update is you use conda