[
https://issues.apache.org/jira/browse/SPARK-14697?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15246184#comment-15246184
]
Joseph K. Bradley commented on SPARK-14697:
-------------------------------------------
This does point to something which has been bugging me: We don't do a good job
of distinguishing between light vs. heavy operations for constructing vectors
and matrices and for accessing their internals, which is important for cheap
conversions between linear algebra libraries. I just created a separate issue
for this: [SPARK-14707]
I'll close this issue, but thanks for bringing this up.
> mllib DenseMatrix toArray could use the internal values
> -------------------------------------------------------
>
> Key: SPARK-14697
> URL: https://issues.apache.org/jira/browse/SPARK-14697
> Project: Spark
> Issue Type: Bug
> Components: MLlib
> Affects Versions: 1.6.1
> Reporter: Tamas Jambor
> Priority: Minor
>
> When toArray is called from an mllib Densematrix, it calls the function from
> the superclass Matrix, where it creates a new array (and iterates through all
> the values).
> This is not necessary as values are stored in DenseMatrix, in exactly the
> format that this function would require. Why not just call that?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
