Skip to content

python 2.7, openml 060 / datasets / datasets.py assertion, special openml-datasets #375

Closed

Description

@bstrang

For some openml tasks (e.g. tid=3512), datasets.py.get_features_by_type() raises an assertion-error,
"row id should be a str".

Maybe a unicode vs. str type issue - special to python 2.7.?
(Issue does not occure when using python 3.6)

Issue Row-id-unicode-vs-string.zip

Activity

  1. mfeurer commented on Nov 17, 2017

    @mfeurer
    Collaborator

    Thanks a lot for reporting this. For the reference, this is the code to reproduce this issue in python2.7:

    import sys
    import openml
    
    print sys.version_info
    print 'OpenML version: ', openml.__version__
    
    task_id = 3
    print 'task_id', task_id
    
    task = openml.tasks.get_task(task_id)
    indices = task.get_dataset().get_features_by_type(data_type='nominal')
    print
    
    task_id = 3512
    print 'task_id', task_id
    
    task = openml.tasks.get_task(task_id)
    indices = task.get_dataset().get_features_by_type(data_type='nominal')
    print
    

    Do you want to contribute to the package and try to work on a fix for this?

  2. janvanrijn commented on Nov 17, 2017

    @janvanrijn
    Member

    Thanks for reporting this, I created a PR.

  3. added a commit that references this issue on Nov 28, 2017
  4. added this to the pip-release milestone on Mar 19, 2018
  5. mfeurer commented on Mar 19, 2018

    @mfeurer
    Collaborator

    This appears to be fixed since November last year.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions