Optimizing Vanilla Python with Pandas and Numpy

Question

Is there anyway that I could make the function below faster and more optimized with pandas or numpy, the function below adds the sum of seq45 until the elements of it is equivalent or over 10000.The elements that is being added up to seq45 are 3,7,11 in order. The reason why i want to increase the speed is to have have to possibly test 10000 with larger number values like 1000000 and to be processed faster. The answer to this code was gotten from this issue: issue

Code:

Sequence = np.array([3, 7, 11])
seq45= []
for n in itertools.cycle(Sequence):
    seq.append(n)
    if sum(seq45) >= 10000: break
        
print(seq45)

Current Perfromance/ Processing time: 71.9ms

The intelligent answers are the ones below, which take advantage of the mathematics of the situation. You don't really have to go through 100,000 iterations. You know that you will have 100000//(3+7+11) complete sets, so you only have to check the leftovers one by one. That will work instantly even with 1,000,000,000 iterations. — Tim Roberts
– Tim Roberts, Commented Mar 6, 2021 at 6:45

bb1 · Accepted Answer · 2021-03-06 06:41:05Z

1

You can try:

Sequence = np.array([3, 7, 11])
s = Sequence.sum()
tot = 10000

seq = list(Sequence)*(tot//s)
mod = tot%s
for n in seq:
    if mod > 0:
        seq.append(n)
        mod -= n
    else:
        break

Wall time: 57 µs

answered Mar 6, 2021 at 6:41

bb1

7,9332 gold badges11 silver badges26 bronze badges

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

1 Comment

9769953 Over a year ago

Fun fact: testing this and my own answer (which are, as pointed out, not precisely equivalent), shows this answer to be slightly faster than using np.tile. But only for tot up to about a million. For larger tot values, np.tile starts to win (my guess is that list(Sequence) * (tot//s) will be expensive), but still only by a small margin. For some things, Python is well optimised.

9769953 · Accepted Answer · 2021-03-06 06:43:23Z

1

import numpy as np

tot = 10_000
sequence = np.array([3, 7, 11])
n = tot // sequence.sum() + 1
seq45 = np.tile(sequence, n)

Note that this is not exactly equivalent, since it repeats the full sequence, not individual elements, and seq45 may thus be slightly larger.

answered Mar 6, 2021 at 6:43

9769953

12.5k3 gold badges31 silver badges46 bronze badges

Collectives™ on Stack Overflow

Optimizing Vanilla Python with Pandas and Numpy

2 Answers 2

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related