You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently cov(x, false) fails when x is a matrix rather than a vector: you need to use cov(x, 1, false). This is annoying since 1) the docstring appears to imply the both arguments are optional, and 2) it makes it harder to find the replacement for the deprecated cov(x, corrected=false) method.
There might be some cleanup left after the removal of the keyword arguments. However, we should think of this with #19451 in mind. We have very few functions left that defaults to dim=1 so we should consider deprecating the cov(Matrix) method.
Currently
cov(x, false)
fails whenx
is a matrix rather than a vector: you need to usecov(x, 1, false)
. This is annoying since 1) the docstring appears to imply the both arguments are optional, and 2) it makes it harder to find the replacement for the deprecatedcov(x, corrected=false)
method.Is this just an oversight?
Cf. https://discourse.julialang.org/t/what-is-the-current-best-practice-for-population-formula-for-cov/658
The text was updated successfully, but these errors were encountered: