python_list to array error

Question

here my code:

ldata = []
for line in f:
    tmp = line.strip().split(',')
    ldata.append([float(i) for i in tmp[2:]])
print len(ldata),len(ldata[0])
cc = np.array(ldata)
print cc.shape
d = cc*0.25

The result:

4345,560
(4345,)
Traceback (most recent call last):
  File "C:\Users\Che\workspace\recali.py", line 62, in <module>
    d = cc*0.25
TypeError: can't multiply sequence by non-int of type 'float'

The shape of the array should (4345,560), while it shows (4345,)....and can not multiply a float. Why did this happen?

I want convert the list to numpy array and then do later calculation(didn't mention in above, the d = cc*0.25 is just a example~) — Owen
– Owen, Commented Jul 1, 2015 at 10:42
if you're multiplying the entire array by 0.25, try d = [x * 0.25 for x in cc] — Dobz
– Dobz, Commented Jul 1, 2015 at 10:43
@RussellHickey, actually, I also checked the type of cc, which shows <type 'numpy.ndarray'>. I believe a array shoud can multiply a float directly — Owen
– Owen, Commented Jul 1, 2015 at 10:55
one of those lists in your lists of lists is not the same length as the others. If you want numpy.array() to make you a 2d array from a list of lists, they have to be all the same length. — NightShadeQueen
– NightShadeQueen, Commented Jul 1, 2015 at 11:02

NightShadeQueen · Accepted Answer · 2015-07-01 11:00:32Z

Somewhere, in your list of lists, is a list of the wrong length

In [3]: numpy.array([range(i) for i in range(50)])*.25 #these lists are of different length
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-3-29e79fee6349> in <module>()
----> 1 numpy.array([range(i) for i in range(50)])*.25

TypeError: can't multiply sequence by non-int of type 'float'

In [4]: numpy.array([range(50) for i in range(50)])*.25 #these lists are of the same length
Out[4]: 
array([[  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25],
       [  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25],
       [  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25],
       ..., 
       [  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25],
       [  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25],
       [  0.  ,   0.25,   0.5 , ...,  11.75,  12.  ,  12.25]])

Dobz · Accepted Answer · 2015-07-01 11:06:14Z

1

The reason you are getting (4345,) as an output is because the lists inside of the array are not the same length, hence it cannot give you the second number.

For example:
Say the array is [[1,2,3], [4,5,6], [7,8,9,10]]. The shape cannot give you the second digit because it can be both 3 or 4.

As for the d = part you should be using d = [[x * 0.25 for x in c] for c in cc]. It will iterate through each list in the array and return a new list multiplied by 0.25.

edited Jul 1, 2015 at 11:06

answered Jul 1, 2015 at 11:03

Dobz

1,2131 gold badge15 silver badges36 bronze badges

2 Comments

NightShadeQueen Over a year ago

You can totally multiply a numpy array by a float and it'll work.

Dobz Over a year ago

@NightShadeQueen The numpy array isn't holding other numpy arrays, it's holding Python ones.

Naman Sogani · Accepted Answer · 2015-07-01 11:10:27Z

0

The error is that ldata list items is itself a list(ldata is a list of lists).

If you want ldata to be 2-d array, then you can do this

ldata_new = []
for data in ldata:
    cc = np.array(data)
    d = cc*0.25
    ldata_new.append(d)
cc = np.array(ldata_new)

Now cc will be of numpy.ndarray type

edited Jul 1, 2015 at 11:10

answered Jul 1, 2015 at 10:48

Naman Sogani

9691 gold badge8 silver badges28 bronze badges

2 Comments

Owen Over a year ago

what i want is a 2-d array, so I can not do this

Naman Sogani Over a year ago

Ok, then you should do it for individual lists of your ldata list rather than multiplying a 2-d list by a float

Collectives™ on Stack Overflow

python_list to array error

3 Answers 3

Comments

2 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related