Merge pull request #124 from osopardo1/42-table-writes-catalog

Add Table properties to Qbeast

README.md

@@ -92,6 +92,7 @@ export SPARK_HOME=$PWD/spark-3.1.1-bin-hadoop3.2
 ```bash
 $SPARK_HOME/bin/spark-shell \
 --conf spark.sql.extensions=io.qbeast.spark.internal.QbeastSparkSessionExtension \
+--conf spark.sql.catalog.spark_catalog=io.qbeast.spark.internal.sources.catalog.QbeastCatalog \
 --packages io.qbeast:qbeast-spark_2.12:0.2.0,io.delta:delta-core_2.12:1.0.0
 ```
 
@@ -118,6 +119,24 @@ csv_df.write
 	.save(tmp_dir)
 ```
 
+#### SQL Syntax.
+You can create a table with Qbeast with the help of `QbeastCatalog`.
+
+```scala
+spark.sql(
+  "CREATE TABLE student (id INT, name STRING, age INT) " +
+    "USING qbeast OPTIONS ('columnsToIndex'='id')")
+
+```
+
+Use **`INSERT INTO`** to add records to the new table. It will update the index in a **dynamic** fashion when new data is inserted.
+
+```scala
+
+spark.sql("INSERT INTO table student SELECT * FROM visitor_students")
+
+```
+
 ###  3. Load the dataset
 Load the newly indexed dataset.
 
@@ -151,21 +170,6 @@ qbeastTable.getIndexMetrics()
 qbeastTable.analyze()
 ```
 
-The format supports **Spark SQL** syntax. 
-It also updates the index in a **dynamic** fashion when new data is inserted.
-
-```scala
-val newData = Seq(1, 2, 3, 4).toDF("value")
-
-newData.createOrReplaceTempView("newTable")
-
-spark.sql("insert into table myTable select * from newTable")
-
-spark.sql("insert into table myTable (value) values (4)")
-
-
-```
-
 Go to [QbeastTable documentation](./docs/QbeastTable.md) for more detailed information.
 
 # Dependencies and Version Compatibility

build.sbt

@@ -146,6 +146,11 @@ ThisBuild / pomExtra :=
       <name>Pol Santamaria</name>
       <url>https://github.com/polsm91</url>
     </developer>
+    <developer>
+      <id>Adricu8</id>
+      <name>Adria Correas</name>
+      <url>https://github.com/Adricu8</url>
+    </developer>
   </developers>
 
 // Scalafmt settings

core/src/main/scala/io/qbeast/core/model/MetadataManager.scala

@@ -63,4 +63,17 @@ trait MetadataManager[DataSchema, FileDescriptor] {
       knownAnnounced: Set[CubeId],
       oldReplicatedSet: ReplicatedSet): Boolean
 
+  /**
+   * Checks if there's an existing log directory for the table
+   * @param tableID the table ID
+   * @return
+   */
+  def existsLog(tableID: QTableID): Boolean
+
+  /**
+   * Creates an initial log directory
+   * @param tableID
+   */
+  def createLog(tableID: QTableID): Unit
+
 }

src/main/scala/io/qbeast/spark/delta/SparkDeltaMetadataManager.scala

@@ -62,4 +62,23 @@ object SparkDeltaMetadataManager extends MetadataManager[StructType, FileAction]
     diff.nonEmpty
   }
 
+  /**
+   * Checks if there's an existing log directory for the table
+   *
+   * @param tableID the table ID
+   * @return
+   */
+  override def existsLog(tableID: QTableID): Boolean = {
+    loadDeltaQbeastLog(tableID).deltaLog.tableExists
+  }
+
+  /**
+   * Creates an initial log directory
+   *
+   * @param tableID
+   */
+  override def createLog(tableID: QTableID): Unit = {
+    loadDeltaQbeastLog(tableID).deltaLog.createLogDirectory()
+  }
+
 }

src/main/scala/io/qbeast/spark/internal/QbeastOptions.scala

@@ -3,6 +3,7 @@
  */
 package io.qbeast.spark.internal
 
+import io.qbeast.core.model.QTableID
 import io.qbeast.spark.index.ColumnsToIndex
 import org.apache.spark.qbeast.config.DEFAULT_CUBE_SIZE
 import org.apache.spark.sql.AnalysisExceptionFactory
@@ -20,6 +21,7 @@ case class QbeastOptions(columnsToIndex: Seq[String], cubeSize: Int)
 object QbeastOptions {
   val COLUMNS_TO_INDEX = "columnsToIndex"
   val CUBE_SIZE = "cubeSize"
+  val PATH = "path"
 
   private def getColumnsToIndex(options: Map[String, String]): Seq[String] = {
     val encodedColumnsToIndex = options.getOrElse(
@@ -49,4 +51,18 @@ object QbeastOptions {
     QbeastOptions(columnsToIndex, desiredCubeSize)
   }
 
+  def loadTableIDFromParameters(parameters: Map[String, String]): QTableID = {
+    new QTableID(
+      parameters.getOrElse(
+        PATH, {
+          throw AnalysisExceptionFactory.create("'path' is not specified")
+        }))
+  }
+
+  def checkQbeastProperties(parameters: Map[String, String]): Unit = {
+    require(
+      parameters.contains("columnsToIndex") || parameters.contains("columnstoindex"),
+      throw AnalysisExceptionFactory.create("'columnsToIndex is not specified"))
+  }
+
 }

src/main/scala/io/qbeast/spark/internal/QbeastSparkSessionExtension.scala

@@ -4,7 +4,7 @@
 package io.qbeast.spark.internal
 
 import io.delta.sql.DeltaSparkSessionExtension
-import io.qbeast.spark.internal.rules.{SampleRule}
+import io.qbeast.spark.internal.rules.{QbeastAnalysis, SampleRule, SaveAsTableRule}
 import org.apache.spark.sql.SparkSessionExtensions
 
 /**
@@ -16,9 +16,17 @@ class QbeastSparkSessionExtension extends DeltaSparkSessionExtension {
 
     super.apply(extensions)
 
+    extensions.injectResolutionRule { session =>
+      new QbeastAnalysis(session)
+    }
+
     extensions.injectOptimizerRule { session =>
       new SampleRule(session)
     }
+
+    extensions.injectOptimizerRule { session =>
+      new SaveAsTableRule(session)
+    }
   }
 
 }

src/main/scala/io/qbeast/spark/internal/rules/QbeastAnalysis.scala

@@ -0,0 +1,41 @@
+/*
+ * Copyright 2021 Qbeast Analytics, S.L.
+ */
+package io.qbeast.spark.internal.rules
+
+import io.qbeast.spark.internal.sources.v2.QbeastTableImpl
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.plans.logical.LogicalPlan
+import org.apache.spark.sql.catalyst.rules.Rule
+import org.apache.spark.sql.execution.datasources.LogicalRelation
+import org.apache.spark.sql.execution.datasources.v2.DataSourceV2Relation
+
+/**
+ * Analyzes and resolves the Spark Plan before Optimization
+ * @param spark the SparkSession
+ */
+class QbeastAnalysis(spark: SparkSession) extends Rule[LogicalPlan] {
+
+  /**
+   * Returns the V1Relation from a V2Relation
+   * @param dataSourceV2Relation the V2Relation
+   * @param table the underlying table
+   * @return the LogicalRelation
+   */
+  private def toV1Relation(
+      dataSourceV2Relation: DataSourceV2Relation,
+      table: QbeastTableImpl): LogicalRelation = {
+
+    val underlyingRelation = table.toBaseRelation
+    LogicalRelation(underlyingRelation, dataSourceV2Relation.output, None, isStreaming = false)
+
+  }
+
+  override def apply(plan: LogicalPlan): LogicalPlan = plan transformDown {
+    // This rule is a hack to return a V1 relation for reading
+    // Because we didn't implemented SupportsRead on QbeastTableImpl yet
+    case v2Relation @ DataSourceV2Relation(t: QbeastTableImpl, _, _, _, _) =>
+      toV1Relation(v2Relation, t)
+  }
+
+}

src/main/scala/io/qbeast/spark/internal/rules/SaveAsTableRule.scala

@@ -0,0 +1,39 @@
+/*
+ * Copyright 2021 Qbeast Analytics, S.L.
+ */
+package io.qbeast.spark.internal.rules
+
+import io.qbeast.spark.internal.sources.catalog.QbeastCatalogUtils.isQbeastProvider
+import org.apache.spark.internal.Logging
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.plans.logical.{
+  CreateTableAsSelect,
+  LogicalPlan,
+  ReplaceTableAsSelect
+}
+import org.apache.spark.sql.catalyst.rules.Rule
+
+/**
+ * Rule class that enforces to pass all the write options to the Table Implementation
+ * @param spark the SparkSession
+ */
+class SaveAsTableRule(spark: SparkSession) extends Rule[LogicalPlan] with Logging {
+
+  override def apply(plan: LogicalPlan): LogicalPlan = {
+    // When CreateTableAsSelect statement is in place for qbeast
+    // We need to pass the writeOptions as properties to the creation of the table
+    // to make sure columnsToIndex is present
+    plan transformDown {
+      case saveAsSelect: CreateTableAsSelect if isQbeastProvider(saveAsSelect.properties) =>
+        val options = saveAsSelect.writeOptions
+        val finalProperties = saveAsSelect.properties ++ options
+        saveAsSelect.copy(properties = finalProperties)
+      case replaceAsSelect: ReplaceTableAsSelect
+          if isQbeastProvider(replaceAsSelect.properties) =>
+        val options = replaceAsSelect.writeOptions
+        val finalProperties = replaceAsSelect.properties ++ options
+        replaceAsSelect.copy(properties = finalProperties)
+    }
+  }
+
+}

src/main/scala/io/qbeast/spark/internal/sources/QbeastBaseRelation.scala

@@ -3,33 +3,25 @@
  */
 package io.qbeast.spark.internal.sources
 
-import io.qbeast.core.model.{QTableID}
 import org.apache.spark.sql.sources.BaseRelation
 import org.apache.spark.sql.sources.InsertableRelation
-
-import org.apache.spark.sql.{SQLContext}
-import org.apache.spark.sql.types.{StructType, StructField}
+import org.apache.spark.sql.SQLContext
+import org.apache.spark.sql.types.{StructField, StructType}
 import org.apache.spark.sql.DataFrame
 import org.apache.spark.sql.SparkSession
 import io.qbeast.spark.delta.OTreeIndex
 import org.apache.spark.sql.execution.datasources.HadoopFsRelation
 import io.qbeast.spark.table.IndexedTable
 import io.qbeast.context.QbeastContext
-import org.apache.hadoop.fs.{Path}
-import org.apache.spark.sql.catalyst.catalog.{BucketSpec}
+import org.apache.hadoop.fs.Path
+import org.apache.spark.sql.catalyst.catalog.BucketSpec
 import org.apache.spark.sql.execution.datasources.parquet.ParquetFileFormat
 
 /**
  * Companion object for QbeastBaseRelation
  */
 object QbeastBaseRelation {
 
-  /**
-   * Creates a QbeastBaseRelation instance
-   * @param tableID the identifier of the table
-   * @return the QbeastBaseRelation
-   */
-
   /**
    * Returns a HadoopFsRelation that contains all of the data present
    * in the table. This relation will be continually updated
@@ -39,48 +31,72 @@ object QbeastBaseRelation {
    * @param sqlContext the SQLContext
    * @return the HadoopFsRelation
    */
-  def createRelation(sqlContext: SQLContext, table: IndexedTable): BaseRelation = {
+  def createRelation(
+      sqlContext: SQLContext,
+      table: IndexedTable,
+      options: Map[String, String]): BaseRelation = {
 
     val spark = SparkSession.active
     val tableID = table.tableID
     val snapshot = QbeastContext.metadataManager.loadSnapshot(tableID)
     val schema = QbeastContext.metadataManager.loadCurrentSchema(tableID)
-    val revision = snapshot.loadLatestRevision
-    val columnsToIndex = revision.columnTransformers.map(row => row.columnName).mkString(",")
-    val cubeSize = revision.desiredCubeSize
-    val parameters =
-      Map[String, String]("columnsToIndex" -> columnsToIndex, "cubeSize" -> cubeSize.toString())
+    if (snapshot.isInitial) {
+      // If the Table is initial, read empty relation
+      // This could happen if we CREATE/REPLACE TABLE without inserting data
+      // In this case, we use the options variable
+      new HadoopFsRelation(
+        OTreeIndex(spark, new Path(tableID.id)),
+        partitionSchema = StructType(Seq.empty[StructField]),
+        dataSchema = schema,
+        bucketSpec = None,
+        new ParquetFileFormat(),
+        options)(spark) with InsertableRelation {
+        def insert(data: DataFrame, overwrite: Boolean): Unit = {
+          table.save(data, options, append = !overwrite)
+        }
+      }
+    } else {
+      // If the table contains data, initialize it
+      val revision = snapshot.loadLatestRevision
+      val columnsToIndex = revision.columnTransformers.map(row => row.columnName).mkString(",")
+      val cubeSize = revision.desiredCubeSize
+      val parameters =
+        Map[String, String]("columnsToIndex" -> columnsToIndex, "cubeSize" -> cubeSize.toString())
 
-    val path = new Path(tableID.id)
-    val fileIndex = OTreeIndex(spark, path)
-    val bucketSpec: Option[BucketSpec] = None
-    val file = new ParquetFileFormat()
+      val path = new Path(tableID.id)
+      val fileIndex = OTreeIndex(spark, path)
+      val bucketSpec: Option[BucketSpec] = None
+      val file = new ParquetFileFormat()
 
-    new HadoopFsRelation(
-      fileIndex,
-      partitionSchema = StructType(Seq.empty[StructField]),
-      dataSchema = schema,
-      bucketSpec = bucketSpec,
-      file,
-      parameters)(spark) with InsertableRelation {
-      def insert(data: DataFrame, overwrite: Boolean): Unit = {
-        table.save(data, parameters, append = !overwrite)
+      new HadoopFsRelation(
+        fileIndex,
+        partitionSchema = StructType(Seq.empty[StructField]),
+        dataSchema = schema,
+        bucketSpec = bucketSpec,
+        file,
+        parameters)(spark) with InsertableRelation {
+        def insert(data: DataFrame, overwrite: Boolean): Unit = {
+          table.save(data, parameters, append = !overwrite)
+        }
       }
     }
   }
 
   /**
    * Function that can be called from a QbeastBaseRelation object to create a
    * new QbeastBaseRelation with a new tableID.
-   * @param tableID the identifier of the table
    * @param indexedTable the indexed table
    * @return BaseRelation for the new table in Qbeast format
    */
-  def forQbeastTable(tableID: QTableID, indexedTable: IndexedTable): BaseRelation = {
+  def forQbeastTable(indexedTable: IndexedTable): BaseRelation = {
+    forQbeastTableWithOptions(indexedTable, Map.empty)
+  }
 
+  def forQbeastTableWithOptions(
+      indexedTable: IndexedTable,
+      withOptions: Map[String, String]): BaseRelation = {
     val spark = SparkSession.active
-    createRelation(spark.sqlContext, indexedTable)
-
+    createRelation(spark.sqlContext, indexedTable, withOptions)
   }
 
 }