[
https://issues.apache.org/jira/browse/FLINK-3996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15307636#comment-15307636
]
ASF GitHub Bot commented on FLINK-3996:
---------------------------------------
Github user danielblazevski commented on a diff in the pull request:
https://github.com/apache/flink/pull/2052#discussion_r65170092
--- Diff:
flink-libraries/flink-ml/src/main/scala/org/apache/flink/ml/math/SparseVector.scala
---
@@ -175,10 +176,21 @@ case class SparseVector(
java.util.Arrays.binarySearch(indices, 0, indices.length, index)
}
+
+ def + (other: Vector): Vector = (this.asBreeze +
other.asBreeze).fromBreeze
+
+ def - (other: Vector): Vector = (this.asBreeze -
other.asBreeze).fromBreeze
+
+ def * (scalar: Double): Vector = (scalar * this.asBreeze).fromBreeze
+
--- End diff --
Remark: when I put `+` and `-` in the `Vector` Trait to avoid duplication
of code, it works fine, but not when I add the `*` method to the `Vector`
Trait. I get errors even when using `+` when I add the `*` method in `Vector`,
namely:
```
Error:(86, 68) could not find implicit value for evidence parameter of type
org.apache.flink.ml.math.BreezeVectorConverter[T]
def + (other: Vector): Vector = (this.asBreeze +
other.asBreeze).fromBreeze
^
```
This error does not appear if I have `+` and `-` in Vector.scala and keep
`*` duplicated in both DenseVector.scala and SparseVector.scala. Found it
weird then to have `+` and `-` in Vector.scala and `*` separate. This is why I
decided to duplicate the code in DenseVector and SparseVector. Happy to hear
if anyone knows of a way to avoid this.
> Add addition, subtraction and multiply by scalar to DenseVector.scala and
> SparseVector.scala
> --------------------------------------------------------------------------------------------
>
> Key: FLINK-3996
> URL: https://issues.apache.org/jira/browse/FLINK-3996
> Project: Flink
> Issue Type: Improvement
> Reporter: Daniel Blazevski
> Assignee: Daniel Blazevski
> Priority: Minor
>
> Small change to add vector operations. With this small change, can now do
> things like:
> val v1 = DenseVector(0.1, 0.1)
> val v2 = DenseVector(0.2, 0.2)
> val v3 = v1 + v2
> instead of what is now has to be done:
> val v1 = DenseVector(0.1, 0.1)
> val v2 = DenseVector(0.2, 0.2)
> val v3 = (v1.asBreeze + v2.asBreeze).fromBreeze
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
