CODEWITHSUNDEEP
pythonpackageversions
defferl
defferl

Reputation: 3

How to manage partly incompatible package versions for Python?

On my search for a modern programming language I found Python and immediately fell in love with it. However, there are some issues and for me the worst is Python version combined with packages.

I started using Python 3.3 and soon found that version numbers are increasing comparatively quickly. Trying some new ones (3.4.3, 3.50b1 and 32bit 3.50b3) I also realized that packages are not necessarily compatible after an upgrade. Meanwhile I am pretty confused and switch back and forth between different versions depending on my needs.

I mainly need to

  1. import data from excel files (csv and text files are no problem) and
  2. apply numerical methods on that data.

Therefore I would like to use openpyxl, numpy, or pandas as an alternative for both.

For coding I use iep and PyCharm on a 64 bit Windows 7. Numpy turned out to be tricky on Win7, but I could install it for Python 3.4.3. It also shows up in PyCharms package list for one of the 32 bit 3.5, but it does not work (some dll missing). Openpyxl instead only works on 3.5, so I cannot use openpyxl and numpy at the same time.

Do you experience such problems too (according to search engines, such issues seem to be common, working solutions seem to be less frequent)? If yes, is there a way how to deal with that properly?

Upvotes: 0

Views: 1217

Answers (1)

Nikita
Nikita

Reputation: 6331

In Python world to separate package versions for different projects tools like "pyvenv" and "virtualenv" are usually used.

Also "conda" environments are used, but less common. Though for data processing one thing suggested is "Anaconda", see https://www.continuum.io/downloads.

Upvotes: 2

Related Questions