Deprecated API

Contents

• Deprecated Interfaces
• Deprecated Classes
• Deprecated Methods

Deprecated Interfaces

Interface and Description
org.apache.spark.AccumulableParam use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam use AccumulatorV2. Since 2.0.0.

Deprecated Classes

Class and Description
org.apache.spark.Accumulable use AccumulatorV2. Since 2.0.0.
org.apache.spark.Accumulator use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam.DoubleAccumulatorParam$ use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam.FloatAccumulatorParam$ use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam.IntAccumulatorParam$ use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam.LongAccumulatorParam$ use AccumulatorV2. Since 2.0.0.
org.apache.spark.AccumulatorParam.StringAccumulatorParam$ use AccumulatorV2. Since 2.0.0.
org.apache.spark.sql.hive.HiveContext Use SparkSession.builder.enableHiveSupport instead. Since 2.0.0.
org.apache.spark.mllib.regression.LassoWithSGD Use ml.regression.LinearRegression with elasticNetParam = 1.0. Note the default regParam is 0.01 for LassoWithSGD, but is 0.0 for LinearRegression. Since 2.0.0.
org.apache.spark.mllib.regression.LinearRegressionWithSGD Use ml.regression.LinearRegression or LBFGS. Since 2.0.0.
org.apache.spark.mllib.classification.LogisticRegressionWithSGD Use ml.classification.LogisticRegression or LogisticRegressionWithLBFGS. Since 2.0.0.
org.apache.spark.mllib.regression.RidgeRegressionWithSGD Use ml.regression.LinearRegression with elasticNetParam = 0.0. Note the default regParam is 0.01 for RidgeRegressionWithSGD, but is 0.0 for LinearRegression. Since 2.0.0.

Deprecated Methods

Method and Description
org.apache.spark.SparkContext.accumulable(R, AccumulableParam) use AccumulatorV2. Since 2.0.0.
org.apache.spark.SparkContext.accumulable(R, String, AccumulableParam) use AccumulatorV2. Since 2.0.0.
org.apache.spark.SparkContext.accumulableCollection(R, Function1>, ClassTag) use AccumulatorV2. Since 2.0.0.
org.apache.spark.SparkContext.accumulator(T, AccumulatorParam) use AccumulatorV2. Since 2.0.0.
org.apache.spark.SparkContext.accumulator(T, String, AccumulatorParam) use AccumulatorV2. Since 2.0.0.
org.apache.spark.scheduler.AccumulableInfo.apply(long, String, Option, String) do not create AccumulableInfo. Since 2.0.0.
org.apache.spark.scheduler.AccumulableInfo.apply(long, String, Option, String, boolean) do not create AccumulableInfo. Since 2.0.0.
org.apache.spark.scheduler.AccumulableInfo.apply(long, String, String) do not create AccumulableInfo. Since 2.0.0.
org.apache.spark.mllib.evaluation.MulticlassMetrics.fMeasure() Use accuracy. Since 2.0.0.
org.apache.spark.TaskContext.isRunningLocally() Local execution was removed, so this always returns false. Since 2.0.0.
org.apache.spark.ml.regression.LinearRegressionSummary.model() The model field is deprecated and will be removed in 2.1.0. Since 2.0.0.
org.apache.spark.ml.classification.RandomForestClassificationModel.numTrees() Use getNumTrees instead. This method will be removed in 2.1.0
org.apache.spark.ml.regression.RandomForestRegressionModel.numTrees() Use getNumTrees instead. This method will be removed in 2.1.0
org.apache.spark.mllib.evaluation.MulticlassMetrics.precision() Use accuracy. Since 2.0.0.
org.apache.spark.mllib.evaluation.MulticlassMetrics.recall() Use accuracy. Since 2.0.0.
org.apache.spark.sql.Dataset.registerTempTable(String) Use createOrReplaceTempView(viewName) instead. Since 2.0.0.
org.apache.spark.sql.Dataset.unionAll(Dataset) use union(). Since 2.0.0.
org.apache.spark.ml.param.Params.validateParams() Will be removed in 2.1.0. All the checks should be merged into transformSchema