Why is py5.random_sample()'s interface different (the replace=True default) from Python's random.sample()?

[villares]

This is not a complaint, I think @hx2A has mentioned the rationale at some point (was it to allow samples bigger than the population?), but I forget and keep being tripped by the replace=True default on py5.random_sample(), because for years I used random.sample() which doesn't repeat elements 😢

• Python's random.sample()
• py5.random_sample()

I think an earlier version of this "conversation" led to the creation of random_permutation() wasn't it? But then, that doesn't allow for a smaller subset... so I either have to remember adding replace=False or I should go back to random.sample().

Edit: Ugh! I can't remember the kwarg :(

[hx2A]

Didn't this derive from random_choice()? I recall that one came first but then we wanted something that would return a list and could return more than one item.

[villares]

I think you are right, and perhaps this is how it works on numpy. Don't you worry about this...