Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/11871#discussion_r57875878
--- Diff:
mllib/src/main/scala/org/apache/spark/ml/feature/StopWordsRemover.scala ---
@@ -123,21 +71,26 @@ class StopWordsRemover(override val uid: String)
/** @group getParam */
def getCaseSensitive: Boolean = $(caseSensitive)
- setDefault(stopWords -> StopWords.English, caseSensitive -> false)
+ setDefault(stopWords -> Array.empty[String], caseSensitive -> false)
override def transform(dataset: DataFrame): DataFrame = {
+ val stopWordsSet = if ($(stopWords).isEmpty) {
+ StopWordsRemover.loadStopWords("english").toSet
+ } else {
+ $(stopWords).toSet
+ }
+
val outputSchema = transformSchema(dataset.schema)
val t = if ($(caseSensitive)) {
- val stopWordsSet = $(stopWords).toSet
- udf { terms: Seq[String] =>
- terms.filter(s => !stopWordsSet.contains(s))
- }
- } else {
- val toLower = (s: String) => if (s != null) s.toLowerCase else s
- val lowerStopWords = $(stopWords).map(toLower(_)).toSet
- udf { terms: Seq[String] =>
- terms.filter(s => !lowerStopWords.contains(toLower(s)))
- }
+ udf { terms: Seq[String] =>
+ terms.filter(s => !stopWordsSet.contains(s))
--- End diff --
I think this can be `terms.filterNot(stopWordsSet.contains)`?
It seems like this code path will always pay the cost of making a set out
of the stopwords. It's not huge but wonder if it makes sense to store a ref to
the set once?
---
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]
