Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/4256#discussion_r23739788
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/linalg/distributed/BlockMatrix.scala
---
@@ -182,6 +184,38 @@ class BlockMatrix(
this
}
+ /** Converts to CoordinateMatrix. */
+ def toCoordinateMatrix(): CoordinateMatrix = {
+ val entryRDD = blocks.flatMap { case ((blockRowIndex, blockColIndex),
mat) =>
+ val rowStart = blockRowIndex.toLong * rowsPerBlock
+ val colStart = blockColIndex.toLong * colsPerBlock
+ mat match {
+ case dn: DenseMatrix =>
+ val entryValues = new ArrayBuffer[MatrixEntry](mat.numRows *
mat.numCols)
--- End diff --
Maybe we can use `mat.foreachActive` directly and start with an empty
`ArrayBuffer`. There will be some overhead of allocating new buffers, but it
should be small. The code will get simplified.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
