I have encountered a behavior relative to package versions in different conda environment that I don’t understand and seems illogical. There is probably something fundamental that I do not understand in the behavior.
Basically :

1. I have a conda environment ‘env’ with pandas version 2.0.3
2. I clone this environment with : “conda create –name env_save –clone env
3. I activate ‘env’
4. I update pandas to 2.2.2
5. I switch back to ‘env_save’ (the one with pandas 2.0.3, I assume)
6. I launch jupyter lab
7. In jupyter lab, in one cell, I import pandas and check for pandas.__version__ : result :2.2.2
8. In the next cell, I check !pip list : result: pandas 2.0.3

Is this a bug or is there a possibility that this ‘spillage’ is as intended ? (the behavior persists even after reboot)