"List file" node - how to get only filenames without path?

Oh, nice! Thank you so much!

I want to revisit this thread:

I often run these two nodes (“List FIles”->“URL to File Path”) to get a list of filenames in a folder. However, “URL to File Path” is SLOW when using cloud drive because it verifies that each file exists before breaking the URL down in its Path components. Since I just ran the List files node, I already know that the files exist and I don’t need to spend the extra 5-10 min waiting for “URL to File Path” to re-check everything. Does anyone know of an equivalent for “URL to File Path” that doesn’t perform the extra checking step?

Note: I am aware that I could use Regex split to manually decompose the URL, however I need a solution that is independent of the system (e.g. Windows, linux), so harcoding in specific path syntax (e.g. “” vs. “/”) isn’t what I’m looking for.

Try Cache node after List Files.

Thanks for the suggestion. I put a cache node between the “List Files” and “URL to File Path” and noticed no difference in behavior. I think this is because the “URL to File Path” will still perform file system calls to confirm each file exists regardless of whether the input table is cached.

Dear @joshuahoran,

Did you find any solution for the same? I also facing same issue…

I want to read File Name along with the data in that file.

The only solution I found (but chose not to implement) was to roll my own RegEx Cell Split in order to manually parse the URLs into file paths and file names. The challenge here is that this is a platform specific task and I need my workflows to run across platforms, so a lot of extra logic based on reflection is required – likely making a very brittle end product.

Hi there,

considering platform issues there is Extract System Properties node which gives you platform information and based on it you can apply right logic using Case of If nodes. Shouldn’t be too complicated.

Considering the speed of URL to File Path node I will check it and get back to you.

Br,
Ivan

Hi there,

to come back to this one. Seems URL to File Path node needs to perform those system calls. Just out of curiosity where are those files located?

Additional workaround for platform issues can be done in Column Expressions node as well. If you need an example I can make one :wink:

Br,
Ivan

Thanks ipazin. The slowness of URL to Pile Path comes into play under the following scenario:
image
You read a list of 2000 file URLs from a remote server (e.g. cloud drive or FTP); then you need to extract the file name, the parent directory name, file extension, etc… The URL to File Path does a great, cross-platform-compatible job with this. However, this node has the side-effect of going through each of those 2000 entries and first verifying that the file exists.This is a one-at-a-time loop and can take 10 minutes depending on the connection speed. This is 100% wasted time since we already know that the files exists because we just obtained the URLs from the List Files node.

Thus having a version of URL-to-FilePath where you could turn off the “verify file exists” part of the functionality would be very useful. Your suggestion that we can re-implement this node from scratch is appealing as it would save hours of time over the course of a week, but I don’t know how to do this so it will work across all platforms (Mac, linux, & windows). An example workflow would be helpful!

Thanks.

3 Likes

Hi @joshuahoran ,

Regex Split with this pattern will give you exactly the same output as the URL to File Path node on the “Location” column:
(.*)(?:[\\/])(.*)(?:\.)(.*)

Edited: It does not matter what OS you are using since both \ and / are accepted.

What do you think @ipazin? Do you think this works fine?

:blush:

6 Likes

Hi there @joshuahoran,

I see. Seems this node indeed performs one IO call per row. I will report it and then maybe something can be improved in future releases.

I didn’t mean to re-implement this node from scratch. What I had in mind is to perform Regex Split based on Platform information. Here is workflow example for you to check out:
https://kni.me/w/L0yXIdSKUMQREVbk

If you can apply right logic based on platform in Regex Split node this should do the trick. And should be faster :wink:

Br,
Ivan

3 Likes

Hi there @armingrudd,

If it can work on all three platforms then great! I have never really worked on Mac or Linux so not familiar with paths and thus can’t verify it covers all three platforms…

Br,
Ivan

2 Likes

Windows uses \ in file paths and everything else uses / as far as I know.

Even if there was an OS using * then there would be no problem at all. You have to add * in the square brackets. :slightly_smiling_face:

:blush:

3 Likes

Thanks ipazin. That looks like a straightforward way to arrange the process. I will see about adopting this.

1 Like

This looks like a nice, flexible bit of RegEx. If it can really handle all OS cases, then that would be great! I will try this one and ipazin’s solution and see what works best.

Thank you.

1 Like

for me this is the best solution! Thanks :smiley:

1 Like

With the new file handling framework you can also use the new path expressions in the Column Expressions node to work with file paths. To get the file name simply use the getFileName() expression as shown in this example workflow:

Bye
Tobias

4 Likes

Hi @tobias.koetter , thanks for sharing, it’s good to know these functions.

Alternatively, you can also get the filename using the URL to File Path node. Of course, we’d need to convert the Path to URI first using the Path to URI node:
image

2 Likes

Thanks for sharing.
This is exactly what I needed.

Extract name files row by row of the data.