When using "Copy data" recursively to copy by last modified, what is the difference between "File path in dataset" vs "Wildcard file path"?

Hello Vadim !

Thank you for posting on Microsoft Learn Q&A.

The 1st file path in dataset you are pointing the dataset to one concrete folder and optionally one concrete file name so the Copy activity will list the files that are under that folder andonly if recursively is checked also under its subfolders and it does not evaluate wildcards here.

The 2nd wildcard file path you can put */? in the folder part or the file name so ADF first expands the pattern and then applies the last modified filter to the matched files and it is useful to pick specific branches or file name patterns across many child folders.

If the dataset version has file path in dataset with a file name filled in, ADF looks only for that exact file in the root folder and it won’t pick differently named files that sit in subfolders.

With wildcard file path, ADF matches all files and then filters by last modified so those subfolder files get included.

If you want to copy everything changed in the tree either you leave the dataset file name empty and keep recursively or use wildcard file path with file: * and if needed wildcards in the folder path to target certain branches.