Why iterating and modifying the each element in nested list results in some wrong output?

Question

The immutable and mutable is causing problem.

I was trying to rotate the matrix by 90 degree clockwise without changing the argument passed in function.

def rotate(m):
 rm=[[0]*len(m)]*(len(m))
 print("m:",m)
 print("rm:",rm)
 for j in range(len(rm)):
 for i in range(len(rm)):
 k=m[j][i]
 y=(len(rm)-j-1)
 rm[i][y] = k
 print(i," ",len(rm)-j-1," ",rm[i][(len(rm)-j-1)])
 print("m:",m)
 print("rm:",rm)
 return rm
 print(rotate([[1,2,3],[4,5,6],[7,8,9]]))

I expect the rm should be "[[7, 4, 1], [8, 5, 2], [9, 6, 3]]" but it is "[[9, 6, 3], [9, 6, 3], [9, 6, 3]]" at the end of function.

Answer 1

You are seeing the error, because the line rm=[[0]*len(m)]*(len(m)). Try to change it for:

rm = [[0 for _ in range(len(m))] for _ in range(len(m))]

Note: The above function can be rewritten as:

l = [[1,2,3],[4,5,6],[7,8,9]]
def rotate(lst):
 return [list(val[::-1]) for val in zip(*lst)]
print(rotate(l))

This prints:

[[7, 4, 1], [8, 5, 2], [9, 6, 3]]

• Wilf Rosenbaum: Wow. I think this is something that is important and non-obvious (maybe it is obvious to everyone but me, I dunno). rm1=[[0]*3]*3 (OPs approach) and rm2=[[0 for _ in range(3)] for _ in range(3)] (Andrej's approach) will produce lists rm1 and rm2 that look alike and compare equal, but behave differently. Assigning rm1[0][0]=33 will result in list [[33, 0, 0], [33, 0, 0], [33, 0, 0]] whereas assigning rm2[0][0]=44 will result in list [[44, 0, 0], [0, 0, 0], [0, 0, 0]].