Let's say I have a couple of dictionaries (one per hour) which all have the same keys:
h1={'a':1, 'b':3, 'c':10}
h2={'a':7, 'b':15, 'c':2}
h3={'a':9, 'b':15, 'c':35}
How can I get the sum of absolute values between the next hours for all hours and keys, meaning a sum of the form:
|1-7|+|7-9|+|3-15|+|15-15|+|10-2|+|2-35|
more generally it is something of the form:
where beta is 1, the psts are the keys of our dictionary and we also sum over hours.
Probably better to use pandas than numpy for this.
Pretty sure it would just be
beta = 1
pd.DataFrame([h1, h2, h3]).diff().dropna().abs().mul(beta).sum().sum()
Out: 61.0
Using numpy, you can craft an array, compute the absolute difference (numpy.diff+abs), and sum:
h1={'a':1, 'b':3, 'c':10}
h2={'a':7, 'b':15, 'c':2}
h3={'a':9, 'b':15, 'c':35}
import numpy as np
dicts = [h1, h2, h3]
a = np.array([list(d.values()) for d in dicts])
abs(np.diff(a, axis=0)).sum()
output: 61
With beta:
beta = 1
abs(np.diff(a, axis=0)*beta).sum()