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
I think an actionable thing to have here would be to have a function oindex that works like ndindex() except it treats the input as an outer index, and converts it into an equivalent vectorized (i.e., normal) index. This would be an extension of the numpy ix_ function, except unlike ix_, it would support arrays with dimensions other than 1, and it would support mixing non-array indices (as long as the array indices aren't separated by non-array indices. That I still want to leave unsupported/unimplemented).
The NEP has some links to implementations which could be useful. What actually would be the most useful would be an implementation that we could test against, but I think we'll just have to write a more direct behavioral test. I expect that to be the hardest part of implementing this.
https://numpy.org/neps/nep-0021-advanced-indexing.html
The text was updated successfully, but these errors were encountered: