Why is Python's list comprehension loop order backwards? [duplicate]

Question

>>> my_list = [[[[1, 2, 3], [4, 5, 6], ]]]
>>> [a for d in my_list for c in d for b in c for a in b]
[1, 2, 3, 4, 5, 6]

is equivalent to

>>> my_list = [[[[1, 2, 3], [4, 5, 6], ]]]
>>> new_list = []
>>> for d in my_list:
...     for c in d:
...         for b in c:
...             for a in b:
...                 new_list.append(a)
... print(new_list):
[1, 2, 3, 4, 5, 6]

This syntax seems backwards when read from left-to-right. According to PEP 202, "The form [... for x... for y...] nests, with the last index varying fastest, just like nested for loops." is "the Right One."

It seems that this order (of left-to-right corresponding to outer-to-inner nested for loops) was chosen because that is the order in which nested for loops are written.

However, since the expression part of the list comprehension (a in the above example), corresponds to the expression at the inner-most part of the nested loops (new_list.append(a) in the above example), it seems to me that the for _ in _ closest to this expression should be the same in both cases, i.e. it should be for a in b and on outwards:

>>> my_list = [[[[1, 2, 3], [4, 5, 6], ]]]
>>> [a for a in b for b in c for c in d for d in my_list]
NameError: name 'b' is not defined

so that the fastest-changing loop is closest to the action, so-to-speak. This also lends itself to being read from left-to-right in more logically stepwise fashion.

Is this a common sentiment among users? or does anyone have a good counter-argument as to why the current syntax implementation really is "the Right One"?

Answer 1

Consider:

[leaf for branch in tree for leaf in branch]

It unrolls like:

for branch in tree:
    for leaf in branch:
        yield leaf

If we write it the other way

[leaf for leaf in branch for branch in tree]

It might make more sense in plain English on some level, but a possible counter-argument is that here the name "branch" is used without being (yet) defined.