A Faster way of Directory walking instead of os.listdir ?

Question

Hello SO!

I am trying to improve performance of elfinder , an ajax based file manager(elRTE.ru) .

It uses os.listdir in a recurisve to walk through all directories recursively and having a performance hit (like listing a dir with 3000 + files takes 7 seconds ) ..

I am trying to improve performance for it here is it's walking function:

        for d in os.listdir(path):
            pd = os.path.join(path, d)
            if os.path.isdir(pd) and not os.path.islink(pd) and self.__isAccepted(d):
                tree['dirs'].append(self.__tree(pd))

My questions are :

1. If i change os.walk instead of os.listdir , would it improve performance?
2. how about using dircache.listdir() ? cache WHOLE directory/subdir contents at the initial request and return cache results , if theres no new files uploaded or no changes in file?
3. Is there any other method of Directory walking which is faster?
4. Any Other Server Side file browser which is fast written in python (but i prefer to make this one fast)?

Answer 1

In order:

• I doubt you'll see much of a speed-up between os.walk and os.listdir, since both rely on the underlying filesystem. In fact, I suspect the underlying filesystem is going to have a big effect on the speed of the operation.

• Any cache operation is going to be significantly faster than hitting the filesystem (at least for the second and subsequent checks).

• You could always write some utility (or call a shell command) which generates the list of directories outside of Python, and called that through the subprocess module. But that's a little complicated, and I'd turn to that solution only if the cache turned out to not work for you.

• If you haven't located a file browser on the Cheeseshop, you probably won't find one.

Answer 2

os.path.walk may increase your performance, for two reasons:

1) If you can stop walking before you've walked everything, then indeed it will be faster than listdir, although only noticeable when dealing with large trees

2) If you're listing HUGE directories, then it can be expensive to make the list returned by listdir. (Not true, see alex's comment below)

However, it probably won't make a difference and may in fact be slower, due to the potentially extra overhead incurred by calling your visit function and doing all the extra argument packing and unpacking.

(Really the only way to answer this question is to test it yourself - it should only take a few minutes)

Answer 3

You should measure directly on the machines (OSs, filesystems and caches thereof, etc) of your specific interest -- whether or not os.walk is faster than os.listdir on a specific and totally different machine / OS / FS will tell you very little about performance on yours.

Not sure what you mean by cachedir.listdir -- no standard library module / function by that name. listdir already reads all the directory in at one gulp (as it must sort the results) as does os.walk (as it must separate subdirectories from files). If, depending on your platform, you have a fast way of being notified about file/directory changes, then it's probably worth building the tree up once and editing it incrementally as change notifications come... but it depends on the relative frequency of changes vs requests, which is, again, totally dependent on your specific application circumstances.