listing studies does not show all studies

[sebffischer]

from openml import study

a = study.list_studies(size = 100)
len(a) 

gives me 2 run studies which is wrong

This has nothing to do with the python api because the same thing happens in R as well.

[PGijsbers]

I looked at this, and it seems to be an issue with the state of the database. As a work-around, you can use the filter /status/in_preparation and it should show the other collections. I don't know from the top of my head which connectors allow you to add these filters.

@joaquinvanschoren Apparently all studies in the database are in_preparation (except 2 run studies and 2 task studies). When no status is specified, the REST API will default to filter on active (php). Do you know why these studies were not set to active?

edit: added v2 tag as the problem either independent of the REST API altogether, or can be resolved in the new API, depending on what actually is wrong. Either way it should stay on our radar.

[Abhii242004]

Hello,

This seems to be a confirmed behavior of the underlying OpenML REST API when listing studies/collections.

The Fix: You need to explicitly set the status filter to 'all' in your function call.

a = study.list_studies(size = 100, status = 'all')
len(a)

Reasoning: When the status parameter is omitted, the API defaults to filtering for only status='active'. It appears most studies in the database are currently marked as in_preparation (as of past discussions), so they are filtered out unless you explicitly ask for all statuses.

Hope this helps!

[joaquinvanschoren]

Made a new release to not require status=active by default.
This will be revisited in API v2 anyway.

Will be released soon.