[SPARK-34517] breeze.linalg.SparseVector.equals always return true may bring error.

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 3.0.2
• Component/s: GraphX

Environment

OpenJDK 1.8.0_275

Scala 2.12.10

Hadoop 3.3.0

Spark 3.0.2

• Assignee: Unassigned
• Reporter: MiKou
• Created: 24/Feb/21 08:48
• Updated: 12/Mar/21 19:40

Description

I implemented an iterative graph algorithm with 'breeze.linalg.SparseVector' just like 'Graph.staticParallelPersonalizedPageRank' does. However, the algorithm can not iterate correctly and I finally found that is because 'breeze.linalg.SparseVector.equals' will always return 'true' if two vectors have the same dimension.

 

scala> val zero =  SparseVector.zeros[Double](10)
zero: breeze.linalg.SparseVector[Double] = SparseVector(10)()
scala> val vec = SparseVector(10)((0, 1.0))
vec: breeze.linalg.SparseVector[Double] = SparseVector(10)((0,1.0))
scala> zero == vec
res2: Boolean = true

 

Does that bring the same error for 'Graph.staticParallelPersonalizedPageRank' ?