How to eliminate loops using numpy array to vectorize a function

Question

I have an array of a stock's price like below. My basic idea is to be able to find the first instance where the price is up by at least 10% and then down by 5% from the first instance where it was up by 10% and on, recursively until the end of array.

price_array=np.array([ 4261.48,  4261.48,  4280.56, ..., 45264.9 , 45240.  , 45215.06])

What i want to do is to create a new array(b) by reading the price_array from left to right where;

b[0] = price_array[0]

b[1] = first value in price_array which is >= 1.1*b[0]

b[2] = first value in price_array after the value of b[1] which is <= 0.95*b[1]

b[3] = first value in price_array after the value of b[2] which is >= 1.1*b[2]

b[4] = first value in price_array after the value of b[3] which is <= 0.95*b[3]

.....................................

.....................................

.....................................

.....................................

.....................................

b[n] = first value in price_array after the value of b[n-1] which is.....

How can we do that using numpy in order to eliminate loops, iterations?

Answer 1

You can divide the array from the first element forward by the array until the last element-1 like this:

price_array[1:]/price_array[:-1]

You'll get an array as a result that represents exactly what you wanted.

result[0] = price_array[1]/price_array[0] 
.
.
.
result[n-2] = price_array[n-1]/price_array[n-2]