os.path.basename(file) vs file.split("/")[-1]

Question

I need to extract seq_00034 from a file path like

    file = "/home/user/workspace/data/seq_00034.pkl"

I know 2 ways to achieve it:

method.A

    import os
    seq_name = os.path.basename(file).split(".")[0]

or

method.B

    seq_name = file.split("/")[-1].split(".")[0]

Which is safer/faster?

(taking the cost of import os into account)

Is there a more elegent way to extract seq_name from given path?

Answer 1

It turns out split twice(i.e. Method B) is faster than os.path + split.

They are both significantly faster than using pathlib

speed test:

import os
import pathlib
import time

given_path = "/home/home/user/workspace/data/task_2022_02_xx_xx_xx_xx.pkl"

time1 = time.time()
for _ in range(10000):
    seq_name = given_path.split("/")[-1].split(".")[0]
print(time.time()-time1, 'time of split')


time2 = time.time()
for _ in range(10000):
    seq_name = pathlib.Path(given_path).stem
print(time.time()-time2, 'time of pathlib')


time3 = time.time()
for _ in range(10000):
    seq_name = os.path.basename(given_path).split(".")[0]
print(time.time()-time3, 'time of os.path')

result (on my PC) is:

0.00339508056640625 time of split
0.0355381965637207 time of pathlib
0.005405426025390625 time of os.path

---

if we take the time consumed for importing into account, split twice (i.e. Method B) is still the fastest

(assume the code is only called once)

time1 = time.time()
seq_name = given_path.split("/")[-1].split(".")[0]
print(time.time()-time1, 'time of split')

time2 = time.time()
import pathlib
seq_name = pathlib.Path(given_path).stem
print(time.time()-time2, 'time of pathlib')

time3 = time.time()
import os
seq_name = os.path.basename(given_path).split(".")[0]
print(time.time()-time3, 'time of os.path')

speed test result:

0.000001430511474609375 time of split
0.003416776657104492 time of pathlib
0.0000030994415283203125 time of os.path