I am trying to determine if a file exists in a network folder:
// File name is "\\QWERTY\folder\dir\A123456.TXT"
Path path = Paths.get("\\\\QWERTY\\folder\\dir\\A123456.TXT")
Using NIO Files
:
Files.exists(path) == false
Using File
:
path.toFile().exists() == true
Using File
seems to be the correct one according to our tests. Why does File
work better than Files
?
So, which is it? Can't be both!
But wait, there is also Files.notExists(path)
.
When the network share file actually exists
Files.exists(path): false
Files.notExists(path): false
path.toFile().exists(): true
When the network share file actually does not exist
Files.exists(path): false
Files.notExists(path): true
path.toFile().exists(): false
Another equally insane way of looking at the three results above
boolean exists = Files.exists(path) || !Files.notExists(path)
boolean notExists = Files.notExists(path) || !Files.exists(path)
boolean oldFashionedExists = path.toFile().exists()
:smileyFace:
Environment and Comments
The program is running on a Windows 8.1 Pro 32 bit machine (OS and machine) and checking on a network share from a Windows 2008 R2 (32 bit) machine.
To determine that Files.exists was failed, I installed a WatchService to monitor the folder and saw that the file did exist when Files.exists was checking. I then logged as both methods and found File.exists to be the correct one.
Now, in my code I have the check as Files.exists(path) || path.toFile().exists()
.
Kinda seems stupid to have to do both. Probably could just get away with the later. Just trying to give the engineers over at Oracle the benefit of the doubt, but the whole thing is rather silly that they report different.
Also, I don't care if 'exists' is immediately outdated. I just want to know if the file exists at the instant that we are checking. I've never come across this -- we just spent 30 hours between me and another developer trying to figure out why our programs are not interfacing because of this 'feature'.
Meditate on this a while
File.exists(): Returns true if and only if the file or directory denoted by this abstract pathname exists; false otherwise.
Files.exists(): Returns true if the file exists; false if the file does not exist or its existence cannot be determined.
That cracks me up! "if and only if the file or directory denoted by this abstract pathname exists; false otherwise" is at odds with "true if the file exists; false if the file does not exist or its existence cannot be determined"
So, how still can File.exists
be true if "the existence cannot be determined"? Obviously, the existence can be (and is being) determined by File but not by Files.
As to why there may be a difference between the two, contrast their documentation:
File.exists(): Returns true if and only if the file or directory denoted by this abstract pathname exists; false otherwise.
Files.exists(): Returns true if the file exists; false if the file does not exist or its existence cannot be determined.
That could possibly explain the difference between the two, perhaps the Files
one is having troubles ascertaining the existence of the file.
For example, under Linux, it's possible to set up directory and file permissions in such a way that you can open a file that exists but cannot see that it exists (by taking away read permission on the directory the file is in while leaving the file permissions more open).
As per more of Oracle's documentation, Files.exists()
only returns true
if the file is verified to exist.
A return value of false
does not mean it doesn't exist.
They suggest you use both exists()
and notExists()
to cover the three possibilities, something like:
if (Files.exists(fspec)) {
System.out.println("It exists!");
else if (Files.notExists(fspec)) {
System.out.println("It does not exist!");
else
System.out.println("I have no idea!");
That covers the three possibilities of file state covered in that link above: