Pulling single file from Sharepoint to append to array WITHOUT using path

[dboren]

Hello,

I've been trying for weeks to solve a problem with merging two pdfs. I'm at the point where I can easily fetch one of the two files in my flow because I can always predict what it's filename is and get file content using path, but getting the other is more troublesome because it's filenames are created with a timestamp in case multiple revisions are made over time.

The merge flow is triggered from a button in a PowerApps customized SharePoint form. The form has a a field for customer name and project number, and those two fields are used to dynamically generate filenames for pdfs that are generated by other flows. Since the first of the two pdfs generates with a timestamp in its filename I'm trying to find a way to use filters in a "get files (properties only)" action to only pull the most recent file with the predictable part of the name as a substring (will add flow images to make this clearer). So far though, I'm still only ending up with one file in the final array of files that I feed to the Merge PDF action. I have an "append to array variable" within the apply to each loop that the filtered get files action has to nest within, but it doesn't seem to be having an effect, and I can't get any info from inside that loop when I check the run logs.

(note: the dynamic compose inputs are being fed from the project ID and customer name fields I mentioned earlier).

If anyone has any recommendations I would greatly appreciate it. Also while it's not reflected in this screenshot I have tried setting the Top Count to 1 in the get files action to try and limit it to only the most recent file containing the substring in its name.

[v-xiaochen-msft]

Hi @dboren ,

Do you mean that the file variable only has one object ?

Could you clearify the issue?

If the file variable only has one object, I guess the 'get files' action returned nothing...

Maybe there is no file meet the 'filter query' condition . Please check the flow run history.

Best Regards,

Wearsky

[dboren · Accepted solution]

Correct. Expecting to have 2 files, but it seems that the append action that occurs within that apply to each loop isn't doing so (but since the run logs don't let you see details within those loops its hard to tell where it's breaking).

I actually ended up with what seems to be a successful workaround to this. The other related flow that creates the first pdf is now saving two copies with each run: one with the timestamp and one without, the latter of which is saved over with each new run of that flow. This means I can predict the file name of the latest timestampless pdf and fetch it by name without having to use any loops.