A string file path was read as bytes in pyreadr.read_r()

[gatorxiao]

I was trying to read a Rda file but kept getting the error below:

File "/usr/local/lib/python3.8/site-packages/pyreadr/pyreadr.py", line 65, in read_r

raise PyreadrError("File {0} does not exist!".format(filename_bytes))

pyreadr.custom_errors.PyreadrError: File b'/path/to/file/file.Rda' does not exist!

I am sure that the path I pass to pyreadr.read_r() is a valid string, but it seems that there is always a b in front of the string. I have tried using strip() to avoid any possible spaces but it didn't work.
Any suggestions would be greatly appreciated!

[ofajardo]

The string is transformed to bytes to be passed to the C library, that detail is not relevant.
You need the file to be in local disk. Remote files or files on mounted remote disks wont work. Pkease check that.

[ofajardo]

Another thing to check: the string must be utf-8 encoded. If it is in another encoding transform it to utf-8. It is described in the docs.

[gatorxiao]

Hi Otto, thank you for your reply. Unfortunately, I still have the same issue. I run the script within Docker. The script and the Rda file are in the same container, so it is not a remote one. The string is utf-8 encoded by default.

[ofajardo]

strange, I work in Docker containers and it works without issues. Maybe it is the way your volumes are mounted? if you are working on a mounted volume have you tried moving your file to your home directory?
when you get the error, the path printed in bytes, if you translate them to a string, is it correct?

In any case, unfortunately the opening of the file is happening at the C library level, which only accepts local files (home directory in docker works for me), and nothing can be done to change that.

[ofajardo]

closed because lack of activity