While is was writing another answer on StackOverflow, I encountered this very strange behaviour: rb seems to be slower than r :

LINE_NUMBER = 1001
NEW_LINE_2 = ""
NEW_LINE_3 = "".encode()
def test2():
    with open("temp.txt", "w") as temp:
        temp.write("Foo\n" * 1000)
        temp.write("REPLACE ME!\n")
        temp.write("Bar\n" * 1000)
    with open("temp.txt", "r ") as temp:
        lines = temp.read().split("\n")
        lines[LINE_NUMBER - 1] = NEW_LINE_2
        temp.seek(0)
        temp.write("\n".join(lines))
        temp.truncate()
def test3():
    with open("temp.txt", "wb") as temp:
        temp.write(b"Foo\n" * 1000)
        temp.write(b"REPLACE ME!\n")
        temp.write(b"Bar\n" * 1000)
    with open("temp.txt", "rb ") as temp:
        lines = temp.read().split(b"\n")
        lines[LINE_NUMBER - 1] = NEW_LINE_3
        temp.seek(0)
        temp.write(b"\n".join(lines))
        temp.truncate()
from timeit import repeat
loops = 3_000
count = 1
print(loops * min(repeat("test2()", globals=globals(), repeat=loops, number=count)))
print(loops * min(repeat("test3()", globals=globals(), repeat=loops, number=count)))

Pydroid 3 (Python 3 on Android):

1.5903121093288064
1.754219876602292 # < slower? How?

https://www.online-python.com/

1.2284908443689346
1.0201307013630867 # faster as expected

I thought that decoding and encoding in order to process the file as a string would be slower than processing the bytes itself. Could someone explain me what's going on? I don't understand why it could be slower on Android. Is this a bug maybe?

CodePudding user response：

This is likely hardware/compiler specific. Try running the code on an online compiler (such as: https://www.online-python.com/ ), and results fluctuate or are reversed

CodePudding user response：

r opens a file for both reading and writing. The file pointer is placed at the beginning of the file.

rb opens a file for both reading and writing in binary format. The file pointer is placed at the beginning of the file.