When prototyping code in a Jupyter notebook, I'm often finding myself using .compute() to check output.
Inspired by numpy transpose which uses the attribute .T for .transpose(), is it worth adding such an attribute for .compute()?
Of course to stay Pythonic I wouldn't want this to confuse matters. .c seems like a possible attribute (I realise if _A_ is a matrix than _A^T_ is the mathematical notation for _A_ transpose, but I don't think there is a meaning for the mathematical notation _A^c_?)
@asmith26 I appreciate the thoughts here but I'm inclined not to add an alias -- at least not at the moment. As you say .T is an alias due to historic mathematical notation and while I can see writing .compute many many times would be annoying, I think it's good to be explicit in dask about compute. Others may disagree. Again, thank you for taking the time to write up your thoughts
Thanks @asmith26 for raising this discussion point and your reasoning behind the proposal! I agree with @quasiben that explicit is better in this case though
Thanks for the feedback. I understand the explanations for not doing this for now, hence closing.
Most helpful comment
@asmith26 I appreciate the thoughts here but I'm inclined not to add an alias -- at least not at the moment. As you say
.Tis an alias due to historic mathematical notation and while I can see writing.computemany many times would be annoying, I think it's good to be explicit in dask aboutcompute. Others may disagree. Again, thank you for taking the time to write up your thoughts