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
where the user was trying to compare two plots that should be identical, and the solution was that in pyplot.scatter the np.ma.ravel() does not convert a 1xN matrix to a 1-D array with length N. But if we do np.ravel() instead of np.ma.ravel() it works...
The text was updated successfully, but these errors were encountered:
This would seem more like a bug with numpy than matplotlib. We need to use np.ma.ravel() as opposed to np.ravel() because we explicitly support masked arrays. While we like to pretend that we support Matrix as well, we don't test them and we don't explicitly say that we support them.
I think it would make sense to explicitly say that we do not support the Matrix subclass, so it is clearly up to the user to convert to ndarray when necessary.
Please, refer to this question in StackOverflow:
http://stackoverflow.com/q/17189313/832621
where the user was trying to compare two plots that should be identical, and the solution was that in pyplot.scatter the np.ma.ravel() does not convert a 1xN matrix to a 1-D array with length N. But if we do np.ravel() instead of np.ma.ravel() it works...
The text was updated successfully, but these errors were encountered: