using multiple functions with correlated arrays Numpy Python

Question

The function below is supposed compute either the Uval and Lval functions in accordance to Set and Numbers. For the first two numbers U and 52599 are in relation and L and 52550 are in relation, since the second number has the label L the Lval equation is used. The previous number is function is previous = numbers[1:] and the current function is current = numbers[:-1]. So (52550 - 52599)/52550 * -100 will be the computation for the first two numbers. The equations are supposed to be computed until the end of the Set and Numbers arrays. However the code gives me the error down below, both the Set and Numbers array have the length of 15.

Function:

Set = np.array(['U', 'L', 'U', 'L', 'U', 'L', 'U', 'L', 'U', 'L', 'U', 'L', 'U', 'L', 'U'])
Numbers = np.array([ 52599, 52550, 53598, 336368, 336875, 337466, 338292, 356587, 357474, 357763, 358491, 358659, 359041, 360179, 360286])
Lval = (Numbers[:-1] - Numbers[1:])/Numbers[:-1] * -100
Uval = (Numbers[1:] - Numbers[:-1])/ Numbers[1:] * -100
Numbers * np.where(Set == 'U', Uval, Lval)

Error Output:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-12-3ddad609950b> in <module>
      4 Uval = (Numbers[1:] - Numbers[:-1])/ Numbers[1:] * -100
      5 print(len(Set))
----> 6 Numbers * np.where(Set == 'U', Uval, Lval)

<__array_function__ internals> in where(*args, **kwargs)

ValueError: operands could not be broadcast together with shapes (15,) (14,) (14,)

Since you are subctracting each element from the previous (or from the subsequent), the shape of Uval and Lval is (14, ) not (15, ). You can try to add padding (I mean, decide what is the previous value for the first element and the subsequent value for the last one) or to select Numbers[1:] or Numbers[:-1] (same for Set) in the last computation. I think use padding is the best solution in your case. Computation made with padding will be discarded. — Salvatore Calcagno
– Salvatore Calcagno, Commented Feb 1, 2021 at 21:08
Could you perhaps give an example of padding i am not sure that I completely understand the concept. — tony selcuk
– tony selcuk, Commented Feb 2, 2021 at 0:44
yes. First, I suggest you to read the answer of BLimitless. I think he noticed a crucial point. Then, in order to handle edges, you coud add fake values to Numbers array at the first and the last position. Numbers = np.array([ 1, 52599, 52550, ..., 360179, 360286, 1]) Fake values depend on your context. In this way you are defining a previous value for the first element and a subsequent value for the last one. When you apply where, just discard padding values Numbers[1:-1] * np.where(Set == 'U', Uval, Lval) — Salvatore Calcagno
– Salvatore Calcagno, Commented Feb 2, 2021 at 7:40

BLimitless · Accepted Answer · 2021-02-02 01:42:51Z

your slice notation isn't doing what you think it is:

in: Numbers[:-1]

does not shift elements in the array by 1. It takes all elements up to but not including the last element. The output is:

in: Numbers[:-1] 
out: array([ 52599,  52550,  53598, 336368, 336875, 337466, 338292, 356587,
   357474, 357763, 358491, 358659, 359041, 360179])

That is every element in Numbers except for the last, 360286.

You need to use the np.roll() function to shift elements by one. This will roll the last element to the front though:

in: np.roll(Numbers, 1)
out: array([360286,  52599,  52550,  53598, 336368, 336875, 337466, 338292,
   356587, 357474, 357763, 358491, 358659, 359041, 360179])

so you have to handle the edge case of what to do with the very first and very last numbers, e.g. your arrays are length 15, so you can only get 14 new elements for either L or U.

Collectives™ on Stack Overflow

using multiple functions with correlated arrays Numpy Python

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related