Exchange and purpose of python

Question

Usually we use A[i], A[j] = A[j], A[i]for swap.

Perhaps this will not work. For instance,

A=[1, 2]
A[0], A[A[0]] = A[A[0]], A[0]

will fail. Why does this fail and how is assignment interpreted?

Answer 1

Look at the bytecode for the statement in question:

>>> def f(A):
...   A[0], A[A[0]] = A[A[0]], A[0]
...
>>> dis.dis(f)
2           0 LOAD_FAST                0 (A)
            3 LOAD_FAST                0 (A)
            6 LOAD_CONST               1 (0)
            9 BINARY_SUBSCR
           10 BINARY_SUBSCR
           11 LOAD_FAST                0 (A)
           14 LOAD_CONST               1 (0)
           17 BINARY_SUBSCR
           18 ROT_TWO
           19 LOAD_FAST                0 (A)
           22 LOAD_CONST               1 (0)
           25 STORE_SUBSCR
           26 LOAD_FAST                0 (A)
           29 LOAD_FAST                0 (A)
           32 LOAD_CONST               1 (0)
           35 BINARY_SUBSCR
           36 STORE_SUBSCR
           37 LOAD_CONST               0 (None)
           40 RETURN_VALUE

This works as follows:

• Instructions 0-10 push A [A [0]] onto the value stack, so this 2
• Instructions 11-17 push A [0] into the value stack, now this 2, 1
• Instruction 18 swaps the stack 1, 2
• Instructions 19–25 assign the upper value (2) A [0], leaving the value stack as 1
• Instructions 26-36 will try to assign the value A [A [0]], but A [0] is now 2, so it tries to assign A [2], which is an IndexError.

, , A[A[0]] , A[0].

:

A0 = A[0]
A[0], A[A0] = A[A0], A[0]