Changes to List Files node in 3.3

KNIME Analytics Platform
1

Greetings,

I'm experiencing an issue with the List Files node in 3.3.2 where using RegEx to filter filenames is not interpretting spaces the same as previous versions (3.1, 3.2).

It appears that the filtering is happening solely on the returned URL field (file:// urls with %20 representing spaces) instead of the Location field (unc path with spaces as expected).

Has anyone else encountered this and is this an expected behavior moving forward? Are there any known workarounds to avoid updating existing workflows to account for this change?

I'm using List Files to get filenames from a file share based on UNC path in a Windows environment (both client and server).

 

Thank you.

2

Hi,

Can you please post the RegEx you are using?

Cheers,

Roland

3

I'm using the following regex to filter files with date/timestamps:

[^~].*Active_Workers_for_IT \d{4}-\d{2}-\d{2} \d{2}_\d{2} \w{3}_toimport\.xlsx

which returns no results. Updating it to:

[^~].*Active_Workers_for_IT%20\d{4}-\d{2}-\d{2}%20\d{2}_\d{2}%20\w{3}_toimport\.xlsx
 

includes the files I'm looking for.

 

The node returns two columns "Location" and "URL"

Location formatted: \\<<servername and folders>>\WD_Active_Workers_for_IT 2016-01-04 04_00 CST_toimport.xlsx

URL formatted: file://<<servername and folders>>/WD_Active_Workers_for_IT%202016-01-04%2004_00%20CST_toimport.xlsx

 

Thank you,

Adam

4

Hi Adam,

Thanks for the reply! 

I have tested this and your expression indeed works in KNIME AP 3.2.2, but not in 3.3.2. I have just opened a ticket so our developers will take care of this. For the time being, there are unfortunately no true workarounds - you will have to change your expressions.

Cheers,

Roland

5

Thank you. Very much appreciated.

I will hold off on upgrading for the time being and retest in future releases.

6

I was able to test this today using the July 2017 KNIME 3.4.0 build and this is working as expected. List Files is allowing me to use spaces in my expressions again.

 

Thank you again for opening the ticket and looking into this.