Why doesn't my loop work as a single line?

Question

I have question about using nested single line for loop in python. In particular, I have the followings:

A = [[tmp1[a][b] / tmp2[a] for b in range(0,10)] for a in range(0,20)]

According to Here, the single line for loop is equivalent as

for a in range(0,20):
    for b in range (0, 10):
        A.append(tmp1[a][b] / tmp2[a])

However, python gives me the following error:

AttributeError: 'numpy.ndarray' object has no attribute 'append'.

How should I modify the structure so that I use double for loop instead of single line nested for loop?

Update:

A=[]
for a in range(0,20):
    B = []
    for b in range (0, 10):
        B.append(tmp1[a][b] / tmp2[a])
   A.append(B)

You did not *initialize* `A` properly... You forgot the `A = []` part. — Willem Van Onsem, Feb 21 '17 at 19:15
It is not a single-line `for` loop; it is a list comprehension. They both simply use the same keyword. — chepner, Feb 21 '17 at 19:15
You're confusing two very similar lines of of code - the first builds a list of lists, the second a single list: `A = [[tmp1[a][b] / tmp2[a] for b in range(0,10)] for a in range(0,20)]` is not the same as `A = [tmp1[a][b] / tmp2[a] for b in range(0,10) for a in range(0,20)]` — Eric, Feb 21 '17 at 19:35
@Eric, please see my update. Do you agree that the update is equivalent to `A = [[tmp1[a][b] / tmp2[a] for b in range(0,10)] for a in range(0,20)]` — wrek, Feb 21 '17 at 19:42

Dadep · Accepted Answer · 2017-02-21T19:46:55.463

1

if you do :

for a in range(0,20):
    for b in range (0, 10):
        A.append(tmp1[a][b] / tmp2[a])

A should be a list to have the attribute append:

A=[]
for a in range(0,20):
    for b in range (0, 10):
        A.append(tmp1[a][b] / tmp2[a])

This method is different it's a list comprehension (a way to create the list) :

A = [tmp1[a][b] / tmp2[a] for a in range(0,20) for b in range(0,10)]

edited Feb 21 '17 at 19:46

answered Feb 21 '17 at 19:19

Dadep

But I do notice a difference though – wrek Feb 21 '17 at 19:34
you have different results ? – Dadep Feb 21 '17 at 19:38
Yes, I think you should go: – wrek Feb 21 '17 at 19:40
`A=[] for a in range(0,20): B = [] for b in range (0, 10): B.append(tmp1[a][b] / tmp2[a]) A.append(B)` – wrek Feb 21 '17 at 19:40
where `tmp1` and `tmp2` come from ? – Dadep Feb 21 '17 at 19:42
Please see Eric's comment – wrek Feb 21 '17 at 19:43
yes sure, sorry didn't read well, this should work better : `A = [tmp1[a][b] / tmp2[a] for a in range(0,20) for b in range(0,10)]` – Dadep Feb 21 '17 at 19:47

score 1 · Answer 2 · answered Feb 21 '17 at 19:48

There are two ways of setting up your list comprehension

This

A = [tmp1[a][b] / tmp2[a] for a in range(0,20) for b in range(0,10)]

produces a flat list.

This:

B = [[tmp1[a][b] / tmp2[a] for b in range(0,10)] for a in range(0,20)]

produces a nested list.

Be sure to note the swap between the two for statements. In the nested case the slowly changing variable comes last, in the flat case it comes first.

The equivalent loops are

A=[]
for a in range(0,20):
    for b in range (0, 10):
        A.append(tmp1[a][b] / tmp2[a])

and

B=[]
for a in range(0,20):
    inner=[]
    for b in range (0, 10):
        inner.append(tmp1[a][b] / tmp2[a])
    B.append(inner)

2 Answers2