I have a script where a user is prompted to type a filename (of a file that is to be opened), and if the file doesn't exist in the current directory, the user is prompted again. Here is the short version:
file = input("Type filename: ")
...
try:
fileContent = open(filename, "r")
...
except FileNotFoundError:
...
When I tested my script on my MacOS X in Python 3.3x it worked perfectly fine when I type the wrong filename on purpose (it executes the suite under "expect").
However, when I wanted to run my code on a Windows computer in Python 3.2x, I get an error that says that "FileNotFoundError" is not defined. So, Python 3.2 on Windows thinks "FileNotFoundError" is a variable and the programs quits with an error.
I figured out that Python 3.2 on Windows throws an "IOError" if the input filename is not valid. I tested it on my Linux machine in Python 2.7, and it's also an IOError.
My problem is now, that the code with
except "FileNotFoundError":
won't run on Windows's Python 3.2, but if I change it to
except "IOError":
it won't work on my Mac anymore.
How could I work around it? The only way I can think of is to use just
except
, which I usually don't want.
This strikes me as better than a simple
except:
, but I'm not sure if it is the best solution:you can catch 2 errors at the same time
except (FileNotFoundError, IOError):
I didn't realize that is what you were asking. I hope there is a more eloquent solution then to manually inspect
except error_to_catch
cwallenpoole does this conditional more eloquently in his answer
(error_to_catch = getattr(__builtins__,'FileNotFoundError', IOError))
So to exactly catch only when a file is not found, I do:
In 3.3,
IOError
became an alias forOSError
, andFileNotFoundError
is a subclass ofOSError
. So you might tryThis will cast a pretty wide net, and you can't assume that the exception is "file not found" without inspecting
e.errno
, but it may cover your use case.PEP 3151 discusses the rationale for the change in detail.