[SPARK-38239] AttributeError: 'LogisticRegressionModel' object has no attribute '_call_java'

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.4.0, 3.0.0, 3.1.0, 3.2.0, 3.3.0
• Fix Version/s: 3.3.0
• Component/s: MLlib, PySpark
• Assignee: Maciej Szymkiewicz
• Reporter: Maciej Szymkiewicz
• Created: 17/Feb/22 12:44
• Updated: 06/Mar/22 16:56
• Resolved: 06/Mar/22 16:56

Description

Trying to invoke __repr__ on pyspark.mllib.classification.LogisticRegressionModel leads to AttributeError:

>>> type(model)
<class 'pyspark.mllib.classification.LogisticRegressionModel'>
>>> model
Traceback (most recent call last):
  File /path/to/python3.9/site-packages/IPython/core/formatters.py:698 in __call__
    return repr(obj)
  File /path/to/spark/python/pyspark/mllib/classification.py:281 in __repr__
    return self._call_java("toString")
AttributeError: 'LogisticRegressionModel' object has no attribute '_call_java'

This problem was introduced SPARK-14712, where the method was added, with the same implementation, for both ml and mllib.

Issue Links

• links to: [Github] Pull Request #35554 (zero323)
• links to: [Github] Pull Request #35554 (zero323)