Github user thvasilo commented on a diff in the pull request:
https://github.com/apache/flink/pull/2740#discussion_r87192044
--- Diff:
flink-libraries/flink-ml/src/main/scala/org/apache/flink/ml/preprocessing/StringIndexer.scala
---
@@ -0,0 +1,108 @@
+package org.apache.flink.ml.preprocessing
+
+import org.apache.flink.api.scala._
+import org.apache.flink.api.scala.extensions.acceptPartialFunctions
+import org.apache.flink.api.scala.utils._
+import org.apache.flink.ml.common.{Parameter, ParameterMap}
+import org.apache.flink.ml.pipeline.{FitOperation,
TransformDataSetOperation, Transformer}
+import org.apache.flink.ml.preprocessing.StringIndexer.HandleInvalid
+
+import scala.collection.immutable.Seq
+
+/**
+ * String Indexer
+ */
+class StringIndexer extends Transformer[StringIndexer] {
+
+ private[preprocessing] var metricsOption: Option[DataSet[(String,
Long)]] = None
+
+
+ def setHandleInvalid(value: String): this.type ={
+ parameters.add( HandleInvalid, value )
+ this
+ }
+
+}
+
+object StringIndexer {
+
+ case object HandleInvalid extends Parameter[String] {
+ val defaultValue: Option[String] = Some( "skip" )
+ }
+
+ // ==================================== Factory methods
==========================================
+
+ def apply(): StringIndexer ={
+ new StringIndexer( )
+ }
+
+ // ====================================== Operations
=============================================
+
+ /**
+ * Trains [[StringIndexer]] by learning the count of each string in the
input DataSet.
+ */
+
+ implicit def fitStringIndexer ={
+ new FitOperation[StringIndexer, String] {
+ def fit(instance: StringIndexer, fitParameters: ParameterMap, input:
DataSet[String]):Unit ={
+ val metrics = extractIndices( input )
+ instance.metricsOption = Some( metrics )
+ }
+ }
+ }
+
+ private def extractIndices(input: DataSet[String]): DataSet[(String,
Long)] = {
+
+ val mapping = input
+ .mapWith( s => (s, 1) )
+ .groupBy( 0 )
+ .reduce( (a, b) => (a._1, a._2 + b._2) )
+ .partitionByRange( 1 )
--- End diff --
Could you explain what is the mapping doing here? If you are trying to sort
shouldn't you be using `.sortPartition()` after the partition?
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
