pandas-profiling (appears to be) a delightful little package that improves on the pd.DataFrame.describe() method. I decided to install it using conda, and, as per the documentation, I input conda install -c conda-forge pandas-profiling on the command line.
Here's where it gets wonky. The current build of pandas-profiling is 2.8.0. This is the text that returned on the command line:
The following packages will be downloaded:
package | build
---------------------------|-----------------
pandas-profiling-1.4.1 | 0 39 KB conda-forge
------------------------------------------------------------
Total: 39 KB
The following NEW packages will be INSTALLED:
pandas-profiling conda-forge/osx-64::pandas-profiling-1.4.1-0
...Version 1.4.1? That doesn't seem right. I'll likely be missing out on some significant functionality. I'd like to upgrade, but, being the neophyte data scientist I am, I'm not familiar enough with conda as a package manager to either figure out why the latest build on conda-forge is 1.4.1 or how to upgrade, either via pip or conda (ideally via conda, for the sake of consistency).
I then investigated the builds with conda search pandas-profiling --info, which revealed the following about the 1.4.1_0 build that appeared to install:
pandas-profiling 1.4.1 0
------------------------
file name : pandas-profiling-1.4.1-0.tar.bz2
name : pandas-profiling
version : 1.4.1
build : 0
build number: 0
size : 39 KB
license : MIT
subdir : osx-64
url : https://conda.anaconda.org/conda-forge/osx-64/pandas-profiling-1.4.1-0.tar.bz2
md5 : 9a23538636bc7bcc05c6e7a90bca3c33
constraints :
- pypy <0a0
dependencies:
- jinja2 >=2.8
- matplotlib >=1.4
- pandas >=0.19
- python >=2.7
- six >=1.9
This was built for Python 2.7, but my conda is managing for version 3.7.6!
Does anyone have any insight into what's going on here and how I can upgrade to the latest build?