Loading of transposed count matrix

Is there an option to pass “transpose” as a parameter to sc.read or sc.read_txt?
It seams cumbersome to load with sc.read(...).transpose()

I don’t find that to be much different than

sc.read(..., transpose=True)

so I’m not sure what the advantage is

I’m not sure as to the internal representation, but it seems weird that we will create an object just so that it’s turned right away… wouldn’t that just take double the memory?

Generally, transpose doesn’t use too much more memory, as most things don’t actually need to be copied. If things are copied, it should only be a temporary increase in memory since the original object would be garbage collected.