How to pull data from GCS without pipelines

Questions
1

Hi,

I am tracking some folders using DVC with google cloud storage. I don’t have any yaml file because I don’t have any experiment.

How can I pull the data from just some folders?

I tried doing dvc pull path_subfolder but this checks for a stage (I can see this with -v). After adding a dummy stage, there with an echo as cmd and the path as deps, the command goes through but no data is present (only in cache).

Thanks for your help.

2

How are you tracking folders from a remote location if you don’t have an import-url stage? How did you get the data in the first place? If you did imported it with import-url in the first place, you should be able to update it as a whole with dvc update.

3

I just did a dvc add folder which created the folder.dvc and then dvc push. So I can do a dvc pull and everything from the folder is getting downloaded, but I want to select some subfolders.

4

I just did a dvc add folder which created the folder.dvc and then dvc push . So I can do a dvc pull and everything from the folder is getting downloaded, but I want to select some subfolders.

Ah, it seems like an issue in dvc 2.0 pull: can't do partial checkouts · Issue #5577 · iterative/dvc · GitHub which is now resolved. When the new bugfix release comes (a couple days), you are going to be able to pull partial paths. Thanks for reporting!

5

Oh ok nice then! Thanks for your help and support :grinning_face_with_smiling_eyes: