Question

I’m using stat to check if a file is still being transferred, but it’s not working. If I copy a large file over the network (talking hundreds of gigabytes, they can take several minutes) and I run the script

from os import stat
from time import sleep
While True:
    stat("thefile.foo")
    sleep(2)

I expect to see either the time or the size changing, but it stays the same. Is there anything else I can use to check if a file is still transferring? This is on a windows server, unfortunately.

I tried this script, to actually read the file:

import sys
finished=False
oldgb=0
while not finished:
f=open(thefile,"rb")
samp=f.read(1)
gb=0
while samp!= b'':
    sys.stdout.flush()
    gb+=1000000000
    f.seek(gb)
    samp=f.read(1)
    print(gb/1000000000,samp)       
f.close()
print(oldgb, gb)
if gb>oldgb:
    oldgb=gb
else:
    finished=True

And noticed that while the file was being copied the bytes that hadn’t been copied yet were initialised to zero, hence the file size not updating. I might use a checksum to check if the file has changed instead.

Answer 1

Ok here’s the hackalicious solution: first I installed sysinternals handle.exe. I use that to see if the files have any open handles, thusly:

import subprocess
from time import sleep
filename="file.foo"
finished = False
while not finished:
    handlesactive=subprocess.check_output("handle.exe")
    activehandles=str(handlesactive).count(filename)
    if activehandles == 0: finished = True
    print("still busy")
    sleep(2)

edit: this works for files being transferred on the machine itself – eg copied from one local file to another, however no file handles are created if a file is being transferred by an external machine. However the mtime of the file is changed in that case (fortunately).

So for my use – in a script to scan a folder and check for new files – I ended up using a combination of checking for handles (for local files) and checking for mtime and size (for external files).