♦ πŸ† 1 min, 🐌 3 min

Append to file from multiple python scripts in parallel

Occasionally we need to append from multiple concurrently running programs to the same file.

For example. Our server is running multiple workers. Each of the workers wants to append data to the same file. python has a built in command to append to a file:

with open("test_file.txt", "a") as file:
file.write(new_line)

using a option we append to the end of the file.

Using a is not the same as opening with w and seeking the end of the file. While opening the file and seeking the end of it another program could write to our file. So we want to append to our test_file.txt.

On some operating systems, opening the file with a guarantees that all your following writes will be appended atomically to the end of the file (even as the file grows by other writes).

Let's test this. We want to run two python scripts in parallel to see if both will write data to the file:

We write script user_a.py:

from datetime import datetime
import time

if __name__ == "__main__":

for i in range(10):
time.sleep(1.0)
current_time = datetime.now()
output = f"a {current_time}\n"

with open("test_file.txt", "a") as file:
file.write(output)

and user_b.py:

from datetime import datetime
import time

if __name__ == "__main__":

for i in range(10):
time.sleep(1)
current_time = datetime.now()
output = f"b {current_time}\n"

with open("test_file.txt", "a") as file:
file.write(output)

One important thing. If we don't want to wait for another script to finish before we can write the data we should close the file immediately after we write to it. This way we reduce the "wait" time for other programs to get access to file writing.

Now create the test file: test_file.txt.

Then open three terminal windows at the same location. In one we'll run the:

tail -f test_file.txt

tail is a linux tool that allows us to monitor how data is appended to a file.

In other two terminals run python3 user_a.py and python3 user_b.py. In the terminal where the tail is running you should see the appropriate output:

That's it. We are now writing from multiple processes to the same file.

One important side note. Depending on the size of the writes the content might not be ordered chronologically. But at least we'll write all the data. If you want to read from multiple processes a file that's changing over time you might want to considerer something more advanced then a txt file.

Get notified & read regularly πŸ‘‡