[Python] Conda downgrading NumPy during package update

I have a virtual Conda environment named dev that was created using the following YAML file:

# *** dev.yml ***

name: dev

channels:
- defaults # Check this channel first
- conda-forge # Fallback to conda-forge if packages are not available in defaults

dependencies:
- python==3.12
- numpy>=2.0.1
# ...more libraries without fixed versions


The environment was created without any dependency issues, and I successfully have NumPy at version 2.0.1. However, when I try to update the packages using the command:

conda update --all


I get the following output suggesting that NumPy will be downgraded:

The following packages will be DOWNGRADED:
numpy 2.0.1-py312h2809609_1 --> 1.26.4-py312h2809609_0
numpy-base 2.0.1-py312he1a6c75_1 --> 1.26.4-py312he1a6c75_0


Why is Conda trying to downgrade NumPy during the update? I understand that dependencies might conflict during an update, but I thought this would be caught during the initial environment creation which runs without dependency warnings.

Is there a way to conditionally update packages where dependencies are solvable while keeping the current NumPy version?

Continue reading...