Github user yinxusen commented on the pull request:
https://github.com/apache/spark/pull/12604#issuecomment-214852153
@jkbradley Find another bug: For LogisticRegression in PySpark, if you
write it then reload it, the result is not identical with the previous one.
Because some params in `ml.LogisticRegression` are defined with default value,
while those params in `pyspark.ml.LogisticRegression` have no default value.
With `JavaParams`, we're going to set default values for those params. So it
will fail the following test:
```python
def _compare_param(self, m1, m2, param):
"""
Compare 2 ML params, assert they have the same param.
"""
# Prevent key not found error in case of some param neither in
paramMap and
# defaultParamMap.
if m1.isDefined(param):
self.assertEqual(m1.getOrDefault(param), m2.getOrDefault(param))
self.assertEqual(param.parent, m2.getParam(param.name).parent)
else:
# If m1 is not defined param, then m2 should not, too.
self.assertEqual(m2.isDefined(m2.getParam(param.name)), False)
```
How about we setup a new JIRA to check the default values?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
