Merge pull request #133 from RedisLabs/fix-issue-132

fix issue #132: scala.MatchError on dataframe read

Author: gkorland

src/main/scala/org/apache/spark/sql/redis/RedisSourceRelation.scala

@@ -157,13 +157,13 @@ class RedisSourceRelation(override val sqlContext: SQLContext,
         new GenericRow(Array[Any]())
       }
     } else {
-      val filteredSchema = {
-        val requiredColumnsSet = Set(requiredColumns: _*)
-        val filteredFields = schema.fields
-          .filter { f =>
-            requiredColumnsSet.contains(f.name)
-          }
-        StructType(filteredFields)
+      // filter schema columns, it should be in the same order as given 'requiredColumns'
+      val requiredSchema = {
+        val fieldsMap = schema.fields.map(f => (f.name, f)).toMap
+        val requiredFields = requiredColumns.map { c =>
+          fieldsMap(c)
+        }
+        StructType(requiredFields)
       }
       val keyType =
         if (persistenceModel == SqlOptionModelBinary) {
@@ -173,12 +173,12 @@ class RedisSourceRelation(override val sqlContext: SQLContext,
         }
       keysRdd.mapPartitions { partition =>
         // grouped iterator to only allocate memory for a portion of rows
-        partition.grouped(iteratorGroupingSize).map { batch =>
+        partition.grouped(iteratorGroupingSize).flatMap { batch =>
           groupKeysByNode(redisConfig.hosts, batch.iterator)
             .flatMap { case (node, keys) =>
-              scanRows(node, keys, keyType, filteredSchema, requiredColumns)
+              scanRows(node, keys, keyType, requiredSchema, requiredColumns)
             }
-        }.flatten
+        }
       }
     }
   }

src/test/scala/com/redislabs/provider/redis/df/HashDataframeSuite.scala

@@ -1,18 +1,21 @@
 package com.redislabs.provider.redis.df
 
 import java.sql.{Date, Timestamp}
+import java.util.UUID
 
 import com.redislabs.provider.redis.toRedisContext
 import com.redislabs.provider.redis.util.Person.{data, _}
 import com.redislabs.provider.redis.util.TestUtils._
 import com.redislabs.provider.redis.util.{EntityId, Person}
 import org.apache.spark.SparkException
-import org.apache.spark.sql.DataFrame
+import org.apache.spark.sql.{DataFrame, Row}
 import org.apache.spark.sql.redis.RedisSourceRelation.tableDataKeyPattern
 import org.apache.spark.sql.redis._
-import org.apache.spark.sql.types._
+import org.apache.spark.sql.types.{StructField, _}
 import org.scalatest.Matchers
 
+import scala.util.Random
+
 /**
   * @author The Viet Nguyen
   */
@@ -295,6 +298,37 @@ trait HashDataframeSuite extends RedisDataframeSuite with Matchers {
     }
   }
 
+  /**
+    * A test case for https://github.com/RedisLabs/spark-redis/issues/132
+    */
+  test("RedisSourceRelation.buildScan columns ordering") {
+    val schema = {
+      StructType(Array(
+        StructField("id", StringType),
+        StructField("int", IntegerType),
+        StructField("float", FloatType),
+        StructField("double", DoubleType),
+        StructField("str", StringType)))
+    }
+
+    val rowsNum = 8
+    val rdd = spark.sparkContext.parallelize(1 to rowsNum, 2).map { _ =>
+      def genStr = UUID.randomUUID().toString
+      def genInt = Random.nextInt()
+      def genDouble = Random.nextDouble()
+      def genFloat = Random.nextFloat()
+      Row.fromSeq(Seq(genStr, genInt, genFloat, genDouble, genStr))
+    }
+
+    val df = spark.createDataFrame(rdd, schema)
+    val tableName = generateTableName("cols-ordering")
+    df.write.format(RedisFormat).option(SqlOptionTableName, tableName).save()
+    val loadedDf = spark.read.format(RedisFormat).option(SqlOptionTableName, tableName).load()
+    loadedDf.schema shouldBe schema
+    loadedDf.collect().length shouldBe rowsNum
+    loadedDf.show()
+  }
+
   def saveMap(tableName: String): Unit = {
     Person.dataMaps.foreach { person =>
       saveMap(tableName, person("name"), person)