[jira] [Commented] (SPARK-18608) Spark ML algorithms that check RDD cache level for internal caching double-cache data

["zhengruifeng (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-18608?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15893596#comment-15893596 ] 

zhengruifeng commented on SPARK-18608:
--------------------------------------

[~mlnick] [~yuhaoyan] [~srowen] I think if we use {{train(dataset: Dataset[_], handlePersistence: Boolean)}} instead of {{train(dataset: Dataset[_])}} may result in extra problems for external implementers, because the existing external algorithms overriding {{Predictor.train}} will not work. 

I think we can do it in another way:
{code}
abstract class Predictor[
    FeaturesType,
    Learner <: Predictor[FeaturesType, Learner, M],
    M <: PredictionModel[FeaturesType, M]]
  extends Estimator[M] with PredictorParams {

  protected var storageLevel = StorageLevel.NONE // 

  override def fit(dataset: Dataset[_]): M = {
    storageLevel = dataset.storageLevel
...
  }

  protected def train(dataset: Dataset[_]): M
{code}

so in algorithm implementations we can use the orignial storageLevel of the input dataset.

> Spark ML algorithms that check RDD cache level for internal caching double-cache data
> -------------------------------------------------------------------------------------
>
>                 Key: SPARK-18608
>                 URL: https://issues.apache.org/jira/browse/SPARK-18608
>             Project: Spark
>          Issue Type: Bug
>          Components: ML
>            Reporter: Nick Pentreath
>
> Some algorithms in Spark ML (e.g. {{LogisticRegression}}, {{LinearRegression}}, and I believe now {{KMeans}}) handle persistence internally. They check whether the input dataset is cached, and if not they cache it for performance.
> However, the check is done using {{dataset.rdd.getStorageLevel == NONE}}. This will actually always be true, since even if the dataset itself is cached, the RDD returned by {{dataset.rdd}} will not be cached.
> Hence if the input dataset is cached, the data will end up being cached twice, which is wasteful.
> To see this:
> {code}
> scala> import org.apache.spark.storage.StorageLevel
> import org.apache.spark.storage.StorageLevel
> scala> val df = spark.range(10).toDF("num")
> df: org.apache.spark.sql.DataFrame = [num: bigint]
> scala> df.storageLevel == StorageLevel.NONE
> res0: Boolean = true
> scala> df.persist
> res1: df.type = [num: bigint]
> scala> df.storageLevel == StorageLevel.MEMORY_AND_DISK
> res2: Boolean = true
> scala> df.rdd.getStorageLevel == StorageLevel.MEMORY_AND_DISK
> res3: Boolean = false
> scala> df.rdd.getStorageLevel == StorageLevel.NONE
> res4: Boolean = true
> {code}
> Before SPARK-16063, there was no way to check the storage level of the input {{DataSet}}, but now we can, so the checks should be migrated to use {{dataset.storageLevel}}.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org