Pretty much what the title says. I\'ve read the documentation and I\'ve played with the function for a while now but I can\'t discern what the physical manifestation of this
This is only applicable for the 2D case, but the two coordinates np.unravel_index functions returns in this case are equivalent to doing floor division and applying the modulo function respectively.
for j in range(1,1000):
for i in range(j):
assert(np.unravel_index(i,(987654321,j))==(i//j,i%j))
The first element of the shape array (ie 987654321) is meaningless except to put an upper bound on how large an unraveled linear index can be passed through the function.