Github user yanboliang commented on a diff in the pull request: https://github.com/apache/spark/pull/19156#discussion_r156517313 --- Diff: mllib/src/main/scala/org/apache/spark/ml/stat/Summarizer.scala --- @@ -197,14 +240,14 @@ private[ml] object SummaryBuilderImpl extends Logging { * metrics that need to de computed internally to get the final result. */ private val allMetrics: Seq[(String, Metric, DataType, Seq[ComputeMetric])] = Seq( - ("mean", Mean, arrayDType, Seq(ComputeMean, ComputeWeightSum)), - ("variance", Variance, arrayDType, Seq(ComputeWeightSum, ComputeMean, ComputeM2n)), + ("mean", Mean, vectorUDT, Seq(ComputeMean, ComputeWeightSum)), + ("variance", Variance, vectorUDT, Seq(ComputeWeightSum, ComputeMean, ComputeM2n)), ("count", Count, LongType, Seq()), - ("numNonZeros", NumNonZeros, arrayLType, Seq(ComputeNNZ)), - ("max", Max, arrayDType, Seq(ComputeMax, ComputeNNZ)), - ("min", Min, arrayDType, Seq(ComputeMin, ComputeNNZ)), - ("normL2", NormL2, arrayDType, Seq(ComputeM2)), - ("normL1", NormL1, arrayDType, Seq(ComputeL1)) + ("numNonZeros", NumNonZeros, vectorUDT, Seq(ComputeNNZ)), --- End diff -- In the old ```mllib.stat.MultivariateOnlineSummarizer```, the internal variable is type of ```Array[Long]```, but the return type is ```Vector```. Do you know the impact of using ```Vector``` internal? Thanks.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org