[SPARK-12160][MLLIB] Use SQLContext.getOrCreate in MLlib

jkbradley · mengxr · commit 3e7e05f5ee76 · 2015-12-07T16:37:09.000-08:00
Switched from using SQLContext constructor to using getOrCreate, mainly in model save/load methods. This covers all instances in spark.mllib. There were no uses of the constructor in spark.ml. CC: mengxr yhuai Author: Joseph K. Bradley <joseph@databricks.com> Closes #10161 from jkbradley/mllib-sqlcontext-fix.
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala b/mllib/src/main/scala/org/apache/spark/mllib/api/python/PythonMLLibAPI.scala
@@ -1191,7 +1191,7 @@ private[python] class PythonMLLibAPI extends Serializable {
   def getIndexedRows(indexedRowMatrix: IndexedRowMatrix): DataFrame = {
     // We use DataFrames for serialization of IndexedRows to Python,
     // so return a DataFrame.
-    val sqlContext = new SQLContext(indexedRowMatrix.rows.sparkContext)
+    val sqlContext = SQLContext.getOrCreate(indexedRowMatrix.rows.sparkContext)
     sqlContext.createDataFrame(indexedRowMatrix.rows)
   }
 
@@ -1201,7 +1201,7 @@ private[python] class PythonMLLibAPI extends Serializable {
   def getMatrixEntries(coordinateMatrix: CoordinateMatrix): DataFrame = {
     // We use DataFrames for serialization of MatrixEntry entries to
     // Python, so return a DataFrame.
-    val sqlContext = new SQLContext(coordinateMatrix.entries.sparkContext)
+    val sqlContext = SQLContext.getOrCreate(coordinateMatrix.entries.sparkContext)
     sqlContext.createDataFrame(coordinateMatrix.entries)
   }
 
@@ -1211,7 +1211,7 @@ private[python] class PythonMLLibAPI extends Serializable {
   def getMatrixBlocks(blockMatrix: BlockMatrix): DataFrame = {
     // We use DataFrames for serialization of sub-matrix blocks to
     // Python, so return a DataFrame.
-    val sqlContext = new SQLContext(blockMatrix.blocks.sparkContext)
+    val sqlContext = SQLContext.getOrCreate(blockMatrix.blocks.sparkContext)
     sqlContext.createDataFrame(blockMatrix.blocks)
   }
 }
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala b/mllib/src/main/scala/org/apache/spark/mllib/classification/NaiveBayes.scala
@@ -192,7 +192,7 @@ object NaiveBayesModel extends Loader[NaiveBayesModel] {
         modelType: String)
 
     def save(sc: SparkContext, path: String, data: Data): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // Create JSON metadata.
@@ -208,7 +208,7 @@ object NaiveBayesModel extends Loader[NaiveBayesModel] {
 
     @Since("1.3.0")
     def load(sc: SparkContext, path: String): NaiveBayesModel = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       // Load Parquet data.
       val dataRDD = sqlContext.read.parquet(dataPath(path))
       // Check schema explicitly since erasure makes it hard to use match-case for checking.
@@ -239,7 +239,7 @@ object NaiveBayesModel extends Loader[NaiveBayesModel] {
         theta: Array[Array[Double]])
 
     def save(sc: SparkContext, path: String, data: Data): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // Create JSON metadata.
@@ -254,7 +254,7 @@ object NaiveBayesModel extends Loader[NaiveBayesModel] {
     }
 
     def load(sc: SparkContext, path: String): NaiveBayesModel = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       // Load Parquet data.
       val dataRDD = sqlContext.read.parquet(dataPath(path))
       // Check schema explicitly since erasure makes it hard to use match-case for checking.
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/classification/impl/GLMClassificationModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/classification/impl/GLMClassificationModel.scala
@@ -51,7 +51,7 @@ private[classification] object GLMClassificationModel {
         weights: Vector,
         intercept: Double,
         threshold: Option[Double]): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // Create JSON metadata.
@@ -74,7 +74,7 @@ private[classification] object GLMClassificationModel {
      */
     def loadData(sc: SparkContext, path: String, modelClass: String): Data = {
       val datapath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val dataRDD = sqlContext.read.parquet(datapath)
       val dataArray = dataRDD.select("weights", "intercept", "threshold").take(1)
       assert(dataArray.size == 1, s"Unable to load $modelClass data from: $datapath")
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/clustering/GaussianMixtureModel.scala
@@ -145,7 +145,7 @@ object GaussianMixtureModel extends Loader[GaussianMixtureModel] {
         weights: Array[Double],
         gaussians: Array[MultivariateGaussian]): Unit = {
 
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // Create JSON metadata.
@@ -162,7 +162,7 @@ object GaussianMixtureModel extends Loader[GaussianMixtureModel] {
 
     def load(sc: SparkContext, path: String): GaussianMixtureModel = {
       val dataPath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val dataFrame = sqlContext.read.parquet(dataPath)
       // Check schema explicitly since erasure makes it hard to use match-case for checking.
       Loader.checkSchema[Data](dataFrame.schema)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/clustering/KMeansModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/clustering/KMeansModel.scala
@@ -124,7 +124,7 @@ object KMeansModel extends Loader[KMeansModel] {
     val thisClassName = "org.apache.spark.mllib.clustering.KMeansModel"
 
     def save(sc: SparkContext, model: KMeansModel, path: String): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
       val metadata = compact(render(
         ("class" -> thisClassName) ~ ("version" -> thisFormatVersion) ~ ("k" -> model.k)))
@@ -137,7 +137,7 @@ object KMeansModel extends Loader[KMeansModel] {
 
     def load(sc: SparkContext, path: String): KMeansModel = {
       implicit val formats = DefaultFormats
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val (className, formatVersion, metadata) = Loader.loadMetadata(sc, path)
       assert(className == thisClassName)
       assert(formatVersion == thisFormatVersion)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/clustering/PowerIterationClustering.scala b/mllib/src/main/scala/org/apache/spark/mllib/clustering/PowerIterationClustering.scala
@@ -70,7 +70,7 @@ object PowerIterationClusteringModel extends Loader[PowerIterationClusteringMode
 
     @Since("1.4.0")
     def save(sc: SparkContext, model: PowerIterationClusteringModel, path: String): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       val metadata = compact(render(
@@ -84,7 +84,7 @@ object PowerIterationClusteringModel extends Loader[PowerIterationClusteringMode
     @Since("1.4.0")
     def load(sc: SparkContext, path: String): PowerIterationClusteringModel = {
       implicit val formats = DefaultFormats
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
 
       val (className, formatVersion, metadata) = Loader.loadMetadata(sc, path)
       assert(className == thisClassName)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala b/mllib/src/main/scala/org/apache/spark/mllib/feature/ChiSqSelector.scala
@@ -134,7 +134,7 @@ object ChiSqSelectorModel extends Loader[ChiSqSelectorModel] {
     val thisClassName = "org.apache.spark.mllib.feature.ChiSqSelectorModel"
 
     def save(sc: SparkContext, model: ChiSqSelectorModel, path: String): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
       val metadata = compact(render(
         ("class" -> thisClassName) ~ ("version" -> thisFormatVersion)))
@@ -150,7 +150,7 @@ object ChiSqSelectorModel extends Loader[ChiSqSelectorModel] {
 
     def load(sc: SparkContext, path: String): ChiSqSelectorModel = {
       implicit val formats = DefaultFormats
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val (className, formatVersion, metadata) = Loader.loadMetadata(sc, path)
       assert(className == thisClassName)
       assert(formatVersion == thisFormatVersion)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/feature/Word2Vec.scala b/mllib/src/main/scala/org/apache/spark/mllib/feature/Word2Vec.scala
@@ -587,7 +587,7 @@ object Word2VecModel extends Loader[Word2VecModel] {
 
     def load(sc: SparkContext, path: String): Word2VecModel = {
       val dataPath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val dataFrame = sqlContext.read.parquet(dataPath)
       // Check schema explicitly since erasure makes it hard to use match-case for checking.
       Loader.checkSchema[Data](dataFrame.schema)
@@ -599,7 +599,7 @@ object Word2VecModel extends Loader[Word2VecModel] {
 
     def save(sc: SparkContext, path: String, model: Map[String, Array[Float]]): Unit = {
 
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       val vectorSize = model.values.head.size
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/recommendation/MatrixFactorizationModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/recommendation/MatrixFactorizationModel.scala
@@ -353,7 +353,7 @@ object MatrixFactorizationModel extends Loader[MatrixFactorizationModel] {
      */
     def save(model: MatrixFactorizationModel, path: String): Unit = {
       val sc = model.userFeatures.sparkContext
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
       val metadata = compact(render(
         ("class" -> thisClassName) ~ ("version" -> thisFormatVersion) ~ ("rank" -> model.rank)))
@@ -364,7 +364,7 @@ object MatrixFactorizationModel extends Loader[MatrixFactorizationModel] {
 
     def load(sc: SparkContext, path: String): MatrixFactorizationModel = {
       implicit val formats = DefaultFormats
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val (className, formatVersion, metadata) = loadMetadata(sc, path)
       assert(className == thisClassName)
       assert(formatVersion == thisFormatVersion)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/regression/IsotonicRegression.scala b/mllib/src/main/scala/org/apache/spark/mllib/regression/IsotonicRegression.scala
@@ -185,7 +185,7 @@ object IsotonicRegressionModel extends Loader[IsotonicRegressionModel] {
         boundaries: Array[Double],
         predictions: Array[Double],
         isotonic: Boolean): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
 
       val metadata = compact(render(
         ("class" -> thisClassName) ~ ("version" -> thisFormatVersion) ~
@@ -198,7 +198,7 @@ object IsotonicRegressionModel extends Loader[IsotonicRegressionModel] {
     }
 
     def load(sc: SparkContext, path: String): (Array[Double], Array[Double]) = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val dataRDD = sqlContext.read.parquet(dataPath(path))
 
       checkSchema[Data](dataRDD.schema)
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/regression/impl/GLMRegressionModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/regression/impl/GLMRegressionModel.scala
@@ -47,7 +47,7 @@ private[regression] object GLMRegressionModel {
         modelClass: String,
         weights: Vector,
         intercept: Double): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // Create JSON metadata.
@@ -71,7 +71,7 @@ private[regression] object GLMRegressionModel {
      */
     def loadData(sc: SparkContext, path: String, modelClass: String, numFeatures: Int): Data = {
       val datapath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val dataRDD = sqlContext.read.parquet(datapath)
       val dataArray = dataRDD.select("weights", "intercept").take(1)
       assert(dataArray.size == 1, s"Unable to load $modelClass data from: $datapath")
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/DecisionTreeModel.scala
@@ -201,7 +201,7 @@ object DecisionTreeModel extends Loader[DecisionTreeModel] with Logging {
     }
 
     def save(sc: SparkContext, path: String, model: DecisionTreeModel): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // SPARK-6120: We do a hacky check here so users understand why save() is failing
@@ -242,7 +242,7 @@ object DecisionTreeModel extends Loader[DecisionTreeModel] with Logging {
 
     def load(sc: SparkContext, path: String, algo: String, numNodes: Int): DecisionTreeModel = {
       val datapath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       // Load Parquet data.
       val dataRDD = sqlContext.read.parquet(datapath)
       // Check schema explicitly since erasure makes it hard to use match-case for checking.
diff --git a/mllib/src/main/scala/org/apache/spark/mllib/tree/model/treeEnsembleModels.scala b/mllib/src/main/scala/org/apache/spark/mllib/tree/model/treeEnsembleModels.scala
@@ -408,7 +408,7 @@ private[tree] object TreeEnsembleModel extends Logging {
     case class EnsembleNodeData(treeId: Int, node: NodeData)
 
     def save(sc: SparkContext, path: String, model: TreeEnsembleModel, className: String): Unit = {
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       import sqlContext.implicits._
 
       // SPARK-6120: We do a hacky check here so users understand why save() is failing
@@ -468,7 +468,7 @@ private[tree] object TreeEnsembleModel extends Logging {
         path: String,
         treeAlgo: String): Array[DecisionTreeModel] = {
       val datapath = Loader.dataPath(path)
-      val sqlContext = new SQLContext(sc)
+      val sqlContext = SQLContext.getOrCreate(sc)
       val nodes = sqlContext.read.parquet(datapath).map(NodeData.apply)
       val trees = constructTrees(nodes)
       trees.map(new DecisionTreeModel(_, Algo.fromString(treeAlgo)))

Original file line number	Diff line number	Diff line change
`@@ -1191,7 +1191,7 @@ private[python] class PythonMLLibAPI extends Serializable {`
`1191`	`1191`	`def getIndexedRows(indexedRowMatrix: IndexedRowMatrix): DataFrame = {`
`1192`	`1192`	`// We use DataFrames for serialization of IndexedRows to Python,`
`1193`	`1193`	`// so return a DataFrame.`
`1194`		`- val sqlContext = new SQLContext(indexedRowMatrix.rows.sparkContext)`
	`1194`	`+ val sqlContext = SQLContext.getOrCreate(indexedRowMatrix.rows.sparkContext)`
`1195`	`1195`	`sqlContext.createDataFrame(indexedRowMatrix.rows)`
`1196`	`1196`	`}`
`1197`	`1197`
`@@ -1201,7 +1201,7 @@ private[python] class PythonMLLibAPI extends Serializable {`
`1201`	`1201`	`def getMatrixEntries(coordinateMatrix: CoordinateMatrix): DataFrame = {`
`1202`	`1202`	`// We use DataFrames for serialization of MatrixEntry entries to`
`1203`	`1203`	`// Python, so return a DataFrame.`
`1204`		`- val sqlContext = new SQLContext(coordinateMatrix.entries.sparkContext)`
	`1204`	`+ val sqlContext = SQLContext.getOrCreate(coordinateMatrix.entries.sparkContext)`
`1205`	`1205`	`sqlContext.createDataFrame(coordinateMatrix.entries)`
`1206`	`1206`	`}`
`1207`	`1207`
`@@ -1211,7 +1211,7 @@ private[python] class PythonMLLibAPI extends Serializable {`
`1211`	`1211`	`def getMatrixBlocks(blockMatrix: BlockMatrix): DataFrame = {`
`1212`	`1212`	`// We use DataFrames for serialization of sub-matrix blocks to`
`1213`	`1213`	`// Python, so return a DataFrame.`
`1214`		`- val sqlContext = new SQLContext(blockMatrix.blocks.sparkContext)`
	`1214`	`+ val sqlContext = SQLContext.getOrCreate(blockMatrix.blocks.sparkContext)`
`1215`	`1215`	`sqlContext.createDataFrame(blockMatrix.blocks)`
`1216`	`1216`	`}`
`1217`	`1217`	`}`