Joseph K. Bradley created SPARK-6661:
----------------------------------------
Summary: Python type errors should print type, not object
Key: SPARK-6661
URL: https://issues.apache.org/jira/browse/SPARK-6661
Project: Spark
Issue Type: Bug
Components: MLlib, PySpark
Affects Versions: 1.3.0
Reporter: Joseph K. Bradley
Priority: Minor
In MLlib PySpark, we sometimes test the type of an object and print an error if
the object is of the wrong type. E.g.:
[https://github.com/apache/spark/blob/f084c5de14eb10a6aba82a39e03e7877926ebb9e/python/pyspark/mllib/regression.py#L173]
These checks should print the type, not the actual object. E.g., if the object
cannot be converted to a string, then the check linked above will give a
warning like this:
{code}
TypeError: not all arguments converted during string formatting
{code}
...which is weird for the user.
There may be other places in the codebase where this is an issue, so we need to
check through and verify.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
