Cannot delete dataset associated with flow created automatically

[marcoslbueno]

Description

Hi,

I uploaded a dataset via the python API (https://www.openml.org/d/42453) and after a few days a task (https://www.openml.org/t/233078) was associated to it. I guess the task was created automatically (by a bot, for example). When I try to delete this dataset I get the error 354 'Dataset is in use by other content. Can not be deleted'. Then, I tried to delete the associated task, but I cannot do so (no remove button is shown to me).

Not sure this is a bug, or just the way I'm trying to delete the dataset.

Steps/Code to Reproduce

Expected Results

Actual Results

Versions

[mfeurer]

Hi, the issue is that a dataset on which content was created cannot be deleted any more, as the other content would then be meaningless.

I'm moving this to the OpenML issue tracker as this is an issue with OpenML an not the python API.

[joaquinvanschoren]

@marcoslbueno: Were you logged in? I could simply remove the task (just did).

@mfeurer: Is it possible that there is no task deletion function in the Python API? I couldn't find it?

[marcoslbueno]

@joaquinvanschoren Yes, I was logged in. The way I was trying to delete the task was probably wrong I guess (via browser)? Thanks anyway, I could now delete the dataset.

[LucaUrbinati44]

@marcoslbueno How did you manage to delete a dataset? I don't find any instruction about it. Thank you

[joaquinvanschoren]

@PGijsbers is this supported by the Python API now?

[PGijsbers]

I am currently working on the PR to allow people to delete any of their uploaded OpenML entities, including datasets. I hope to finish it next week.

[LucaUrbinati44]

I found this function by looking at the source code of OpenML (in the documentation there is nothing about it) and it works:

# Remove dataset
openml.utils._delete_entity("data", <your_dataset_id>)

[PGijsbers]

Yes, the PR is just to make documented functions and better error messaging surrounding that function: #1060