4D数组中最小/最大元素的索引

Question

I need to change min/max values of a 4D array along specific axis. The coordinates of min/max along a particular axis are retuned by armax , so for 4D array a have a 3D array.

Now from this 3D array I need to recover the full index of min/max values. I understand the most efficient way to do it is combining results of mgrid and argmin using something like (for min along 1st axis):

mg = np.mgrid[0:Array.shape[0], 0:Array.shape[1], 0:Array.shape[2], 0:Array.shape[3]]

Array[mg[0], np.argmin(Array, axis=1), mg[2], mg[3]] = np.min(Array)

however, this does not seem to work correctly.

Could someone please advise on how to correct the code above to replace the min values of Array along 1st axis with absolute min of Array?

Alternatively, if this is not the fastest way to do it, what is the optimal solution?

Answer 1

Try indexing with:

Array[:, np.argmin(Array, axis=1), :, :] = ....

I think you are trying to set all the values of one 'plane' to a minimum value. : is a slice , specifying all indexes along that axis.

I haven't actually tried to run your example, but I don't think you need to generate mg for this purpose.