Big file upload does not work

[AdrianoW]

Hi Folks,

I have used the example provided to upload a big file. It creates the file in Sharepoint but the contents are not filled.

After some debugging, it seemed to me that the upload operation would start but would never be finished.

So I did the following changes to the code:
file upload_session.py
from:

for piece in read_in_chunks(fh, size=self._chunk_size):
    if f_pos == 0:
        upload_result = files.get_by_url(file_name).start_upload(self._upload_id, piece)
    elif f_pos + len(piece) < st.st_size:
        upload_result = files.get_by_url(file_name).continue_upload(self._upload_id, f_pos, piece)
    else:
        self._target_file = files.get_by_url(file_name).finish_upload(self._upload_id, f_pos, piece)
    f_pos += len(piece)

to

for piece in read_in_chunks(fh, size=self._chunk_size): # when the bytes are over, the loop gets out, without calling the else in the previous code
    if f_pos == 0:
        upload_result = files.get_by_url(file_name).start_upload(self._upload_id, piece)
    elif f_pos + len(piece) < st.st_size:
        upload_result = files.get_by_url(file_name).continue_upload(self._upload_id, f_pos, piece)
    f_pos += len(piece)

# in my opinion this should always be called, as this is what tells sharepoint to close the file
self._target_file = files.get_by_url(file_name).finish_upload(self._upload_id, f_pos, piece)

[vgrem]

Greetings!

thanks for spotting this bug, indeed, due to the recent changes to ClientContext, continue_upload among with finish_upload stopped gets triggered, that's the reason why the file gets uploaded as an empty one.

The good news, the latest commit addresses this particular issue (and it is now part of unit tests).

And since it is not yet published into PyPI, the latest version could be grabbed from GitHub:

pip install git+https://github.com/vgrem/Office365-REST-Python-Client.git

Here is an example on how to upload a file via Session API:

def print_upload_progress(offset):
    print("Uploaded '{0}' bytes...".format(offset))   


if __name__ == '__main__':
    ctx = ClientContext.connect_with_credentials(url, UserCredential(username,password))
    size_1Mb = 1000000
    local_path = "../../tests/data/big_buck_bunny.mp4"
    target_url = "/Shared Documents"
    result_file = ctx.web.get_folder_by_server_relative_url(target_url) \
        .files.create_upload_session(local_path, size_1Mb, print_upload_progress)
    ctx.execute_query()
    print('File {0} has been uploaded successfully'.format(result_file.serverRelativeUrl))