[
https://issues.apache.org/jira/browse/SPARK-9750?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14696046#comment-14696046
]
Joseph K. Bradley commented on SPARK-9750:
------------------------------------------
My understanding is as follows. Does this sound right?
Terminology:
* literal equality: ignores semantics of tranpose, sparsity, etc., and simply
compares objects for exact equality
* semantic equality: equality in linear algebra semantics, e.g., as Breeze does
Status before this PR:
* Python
** Vectors: eq implements literal equality. Default hash.
** Matrices: eq implements semantic equality. Default hash. **-->This is a
bug.**
* Scala
** Vectors: equals() implements semantic equality. hashCode() uses first 16
elements.
** DenseMatrix: equals() implements semantic equality. hashCode() uses all
elements.
** SparseMatrix: default equals(), hashCode().
Thus, there is one bug which we could fix. But it's probably not significant
since it's been in MLlib for a while.
> SparseMatrix should override equals
> -----------------------------------
>
> Key: SPARK-9750
> URL: https://issues.apache.org/jira/browse/SPARK-9750
> Project: Spark
> Issue Type: Bug
> Components: MLlib
> Reporter: Feynman Liang
> Assignee: Feynman Liang
> Priority: Critical
> Fix For: 1.5.0
>
>
> [SparseMatrix|https://github.com/apache/spark/blob/9897cc5e3d6c70f7e45e887e2c6fc24dfa1adada/mllib/src/main/scala/org/apache/spark/mllib/linalg/Matrices.scala#L479]
> should override equals to ensure that two instances of the same matrix are
> equal.
> This implementation should take into account the {{isTransposed}} flag and
> {{values}} may not be in the same order.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
