How to index within array in python for loop?

Question

I've checked some previous questions but can't quite seem to get an answer, though I wouldn't be surprised if this has been asked before. I would like to perform some calculations on elements of an array that I index using a for loop.

I have two arrays:

    S = [[4.21287783e-03 7.83625813e-03 1.42038926e-02 ... 3.15416197e-03
      1.37110355e-03 9.45473448e-04]
     [1.94774282e-03 1.36746081e-03 1.23485391e-03 ... 6.21054272e-04
      5.31808587e-04 1.78796272e-04]
     [1.20601337e-03 2.81822793e-04 6.32125664e-04 ... 2.72966598e-04
      3.88162201e-04 1.89432902e-04]
     ...
     [7.39537451e-05 1.20665168e-04 1.54863119e-04 ... 3.05247233e-04
      2.26473099e-04 1.56650857e-04]
     [9.29507556e-05 6.45091024e-05 9.84829924e-05 ... 3.07827294e-04
      2.33815251e-04 1.52187484e-04]
     [4.66322444e-05 3.16681323e-05 7.08467828e-05 ... 1.44890351e-04
      7.91870831e-05 5.80408583e-05]]

    frames = [ 1  2  3  4  5  6  7  8  9 ]

I iterate through my frames array, but want to iteratively perform a calculation on a single value (indexed with i) from the S array:

    for i in frames:
          np.log(S[:,i])

But I get an out-of-bounds error, ('index 9 is out of bounds for axis 1 with size 9') because i indexes to the end of frames. I tried:

    np.log(S[:,(i-1)])

which didn't work - either because my syntax is wrong or my logic is wrong.

I also tried:

      for i in frames:
          i=i-1
          np.log(S[:,i])

And get the same out of bounds error.

EDIT: I am confident that I can call S in this manner because I do so elsewhere in the script (and can sub in any integer for i and the script runs). My logic about using i as an index is wrong.

hpaulj · Accepted Answer · 2019-04-18 01:39:42Z

2

With the two lists that you define (you write about arrays but a copy-n-paste of your code produces lists):

In [30]: S = [23, 23.3, 34.2, 235, 23.1, 32.1, 23, 75, 4] 
    ...: frames = [1, 2, 3, 4, 5, 6, 7, 8, 9]                                        
In [31]: for i in frames: 
    ...:     print(S[:,i]) 
    ...:                                                                             
TypeError: list indices must be integers or slices, not tuple

You can't use [:,i] indexing with lists.

In [32]: for i in frames: 
    ...:     print(S[i]) 
    ..:                                                                             
23.3
34.2
235
23.1
32.1
23
75
4
---------------------------------------------------------------------------
IndexError: list index out of range

With frames you miss the first element of S, and get an error with the last index. Python indexing starts with 0!

Even if I make a numpy array your indexing is wrong:

In [33]: arr = np.array(S)                                                           
In [34]: for i in frames: 
    ...:     print(arr[:,i]) 
    ...:                                                                             

IndexError: too many indices for array

arr is 1d, shape (9,). You can't use [:,i]` with that.

Do you want to select a part of S (or arr), for example the first the 3 elements?

In [36]: arr[:3]                                                                     
Out[36]: array([23. , 23.3, 34.2])
In [37]: np.log(arr[:3])                                                             
Out[37]: array([3.13549422, 3.14845336, 3.53222564])

[:3] indexes a slice (both for list and arrays)

If the array is 2d, then you can use the [:,i] notation:

In [38]: A = arr.reshape(3,3)                                                        
In [39]: A                                                                           
Out[39]: 
array([[ 23. ,  23.3,  34.2],
       [235. ,  23.1,  32.1],
       [ 23. ,  75. ,   4. ]])
In [40]: A[:,0]                       # first column                                               
Out[40]: array([ 23., 235.,  23.])

edited Apr 18, 2019 at 1:39

answered Apr 18, 2019 at 1:33

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

MeC Over a year ago

Thanks for taking the time to walk through this. I was indeed misidentifying the data structure (now edited). And I've clarified: I want to get the log of just one element of S (indexed with i). How do I avoid the out of bounds error?

hpaulj Over a year ago

Normally the index i is created within bounds. If S has shape (n,) (n elements), frame will be all or some subset of [0...8]. That's what range(n) or np.arange(n) does. In other words, frame is created with full knowledge of array it will index. If an element of frame is too large, you have to decide what that means - is it an error, do you clip it at the maximum allowed, or does it wrap around, or some other magic?

MeC Over a year ago

Got it. Seems like it's a larger issue with my data structure then and I may have been making some incorrect assumptions about calculations I did elsewhere in the script. At least I'm glad to know that my syntax wasn't off. Thanks for your time.

Martin · Accepted Answer · 2019-04-18 00:49:18Z

0

You can remove "9" in frames array, and you will see the result as below:

So now you know the array index starts from 0, not from 1. if you want to fix that, you need to replace the frames array:

frames = [0, 1, 2, 3, 4, 5, 6, 7, 8]

answered Apr 18, 2019 at 0:49

Martin

11 bronze badge

Comments

Alfa Cento · Accepted Answer · 2019-04-18 00:53:48Z

0

You have an extra comma.

Try this:

for i in frames:
      np.log(S[:i])

My test:

$ python3
Python 3.7.2+ (default, Feb 27 2019, 15:41:59) 
[GCC 8.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> S = [23, 23.3, 34.2, 235, 23.1, 32.1, 23, 75, 4]
>>> frames = [1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> for i in frames:
...     print(S[:i])
... 
[23]
[23, 23.3]
[23, 23.3, 34.2]
[23, 23.3, 34.2, 235]
[23, 23.3, 34.2, 235, 23.1]
[23, 23.3, 34.2, 235, 23.1, 32.1]
[23, 23.3, 34.2, 235, 23.1, 32.1, 23]
[23, 23.3, 34.2, 235, 23.1, 32.1, 23, 75]
[23, 23.3, 34.2, 235, 23.1, 32.1, 23, 75, 4]
>>>

answered Apr 18, 2019 at 0:53

Alfa Cento

11 bronze badge

Comments

haofeng · Accepted Answer · 2019-04-18 00:55:08Z

0

Personally, I think you dont need to use frames to index S. You can try this way:

for i in range(S.shape[0])
    np.log(S[i])

edited Apr 18, 2019 at 0:55

answered Apr 18, 2019 at 0:44

haofeng

6711 gold badge8 silver badges22 bronze badges

4 Comments

MeC Over a year ago

'frames' is just a pseudo-array. The actual 'frames' is a much more complicated array drawn from variables defined elsewhere in the script. So unfortunately it's not as easy as just looking at the length of S.

haofeng Over a year ago

Basically, S.shape is (9, ) , You use S[:,i] (2d index) to index 1d array. This is not right I think.

haofeng Over a year ago

I dont know what do you mean about a particular part of the S array. Because if S is the list, it is easy to output a part of S. For example, the first to third part of S is S[0:3], at least without dot",".

MeC Over a year ago

It's not a list, it's a 2D array.

Collectives™ on Stack Overflow

How to index within array in python for loop?

4 Answers 4

3 Comments

Comments

Comments

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

3 Comments

Comments

Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related