Why Is Django Fieldfile Readline() Returning The Hex Version Of A Text File?

Having an odd problem. I have a Django app that opens a file (represented as a Django FieldFile) and reads each row using readline() as below: with file.open(mode='r') as f: row

Solution 1:

This seems very weird. I think you will see the solution immediately after trying following (I will then update my answer or delete it if it really doesn't help to find it, but I'm quite confident)

A assume, that there is some code, that is monkeypatching file.open or the django view function.

What I suggest is:

Start your code with manage.py runserver Ad following code to manage.py (as the very first lines)

import file
print("ID of file.open at manage startup is", id(file.open)

Then add code to your view directly one line above the file.open

print("ID of file.open before opening is", id(file.open)

If both ids are different, then something monkeypatched your open function. If both are the same, then the problem must be somewhere else.

If you don not see the output of these two prints, something might have monkeypatched your view.

If this doesn't work, then try to use open() instead of file.open()

Is there any particular reason you use file.open()

Addendum 1:

So what you sai is, that file is an object instance of a class is it a FileField? In any case can you obtain the name of the file and open it with a normal open() to see whether it is only file.open() that does funny things or whether it is also open() reading it this stange way. Did you just open the file from command line with cat filename (or if under windows with type filename?

If that doesn't work we could add traces to follow each line of the source code that is being executed.

Addendum 2:

Well if you can't try this in a manage.py runserver, what happens if you try to read the file with a manage.py shell?

Just open the shell and type something like:

from <your_application>.models import <YourModel>
entry = <YourModel>.objects.get(id=<idofentry>)
line1 = entry.<filefieldname>.open("r").read().split("\n")[0]
print("line1 = %r" % line1)

If this is still not conclusive, (but only if you can reproduce the issue with the management shell, then create a small file containing the lines.

from <your_application>.models import <YourModel>
entry = <YourModel>.objects.get(id=<idofentry>)
import pdb; pdb.set_trace()
line1 = entry.<filefieldname>.open("r").read().split("\n")[0]
print("line1 = %r" % line1)

And import it from the management shell. The code should enter the debugger and now you can single step through the open function and see whether you end up on sime weird function in some monkeypatch.