[SPARK-26021][SQL][followup] add test for special floating point values

common/unsafe/src/test/java/org/apache/spark/unsafe/PlatformUtilSuite.java

@@ -165,10 +165,14 @@ public void writeMinusZeroIsReplacedWithZero() {
     byte[] floatBytes = new byte[Float.BYTES];
     Platform.putDouble(doubleBytes, Platform.BYTE_ARRAY_OFFSET, -0.0d);
     Platform.putFloat(floatBytes, Platform.BYTE_ARRAY_OFFSET, -0.0f);
-    double doubleFromPlatform = Platform.getDouble(doubleBytes, Platform.BYTE_ARRAY_OFFSET);
-    float floatFromPlatform = Platform.getFloat(floatBytes, Platform.BYTE_ARRAY_OFFSET);
 
-    Assert.assertEquals(Double.doubleToLongBits(0.0d), Double.doubleToLongBits(doubleFromPlatform));
-    Assert.assertEquals(Float.floatToIntBits(0.0f), Float.floatToIntBits(floatFromPlatform));
+    byte[] doubleBytes2 = new byte[Double.BYTES];
+    byte[] floatBytes2 = new byte[Float.BYTES];
+    Platform.putDouble(doubleBytes, Platform.BYTE_ARRAY_OFFSET, 0.0d);
+    Platform.putFloat(floatBytes, Platform.BYTE_ARRAY_OFFSET, 0.0f);
+
+    // Make sure the bytes we write from 0.0 and -0.0 are same.
+    Assert.assertArrayEquals(doubleBytes, doubleBytes2);
+    Assert.assertArrayEquals(floatBytes, floatBytes2);
   }
 }

docs/sql-migration-guide-upgrade.md

@@ -17,14 +17,16 @@ displayTitle: Spark SQL Upgrading Guide
 
   - Since Spark 3.0, the `from_json` functions supports two modes - `PERMISSIVE` and `FAILFAST`. The modes can be set via the `mode` option. The default mode became `PERMISSIVE`. In previous versions, behavior of `from_json` did not conform to either `PERMISSIVE` nor `FAILFAST`, especially in processing of malformed JSON records. For example, the JSON string `{"a" 1}` with the schema `a INT` is converted to `null` by previous versions but Spark 3.0 converts it to `Row(null)`.
 
-  - In Spark version 2.4 and earlier, the `from_json` function produces `null`s for JSON strings and JSON datasource skips the same independetly of its mode if there is no valid root JSON token in its input (` ` for example). Since Spark 3.0, such input is treated as a bad record and handled according to specified mode. For example, in the `PERMISSIVE` mode the ` ` input is converted to `Row(null, null)` if specified schema is `key STRING, value INT`. 
+  - In Spark version 2.4 and earlier, the `from_json` function produces `null`s for JSON strings and JSON datasource skips the same independetly of its mode if there is no valid root JSON token in its input (` ` for example). Since Spark 3.0, such input is treated as a bad record and handled according to specified mode. For example, in the `PERMISSIVE` mode the ` ` input is converted to `Row(null, null)` if specified schema is `key STRING, value INT`.
 
   - The `ADD JAR` command previously returned a result set with the single value 0. It now returns an empty result set.
 
   - In Spark version 2.4 and earlier, users can create map values with map type key via built-in function like `CreateMap`, `MapFromArrays`, etc. Since Spark 3.0, it's not allowed to create map values with map type key with these built-in functions. Users can still read map values with map type key from data source or Java/Scala collections, though they are not very useful.
-  
+
   - In Spark version 2.4 and earlier, `Dataset.groupByKey` results to a grouped dataset with key attribute wrongly named as "value", if the key is non-struct type, e.g. int, string, array, etc. This is counterintuitive and makes the schema of aggregation queries weird. For example, the schema of `ds.groupByKey(...).count()` is `(value, count)`. Since Spark 3.0, we name the grouping attribute to "key". The old behaviour is preserved under a newly added configuration `spark.sql.legacy.dataset.nameNonStructGroupingKeyAsValue` with a default value of `false`.
 
+  - In Spark version 2.4 and earlier, float/double -0.0 is semantically equal to 0.0, but users can still distinguish them via `Dataset.show`, `Dataset.collect` etc. Since Spark 3.0, float/double -0.0 is replaced by 0.0 internally, and users can't distinguish them any more.
+
 ## Upgrading From Spark SQL 2.3 to 2.4
 
   - In Spark version 2.3 and earlier, the second parameter to array_contains function is implicitly promoted to the element type of first array type parameter. This type promotion can be lossy and may cause `array_contains` function to return wrong result. This problem has been addressed in 2.4 by employing a safer type promotion mechanism. This can cause some change in behavior and are illustrated in the table below.

sql/catalyst/src/main/java/org/apache/spark/sql/catalyst/expressions/UnsafeArrayData.java

@@ -334,17 +334,11 @@ public void setLong(int ordinal, long value) {
   }
 
   public void setFloat(int ordinal, float value) {
-    if (Float.isNaN(value)) {
-      value = Float.NaN;
-    }
     assertIndexIsValid(ordinal);
     Platform.putFloat(baseObject, getElementOffset(ordinal, 4), value);
   }
 
   public void setDouble(int ordinal, double value) {
-    if (Double.isNaN(value)) {
-      value = Double.NaN;
-    }
     assertIndexIsValid(ordinal);
     Platform.putDouble(baseObject, getElementOffset(ordinal, 8), value);
   }

sql/core/src/test/scala/org/apache/spark/sql/DatasetPrimitiveSuite.scala

@@ -393,4 +393,33 @@ class DatasetPrimitiveSuite extends QueryTest with SharedSQLContext {
     val ds = spark.createDataset(data)
     checkDataset(ds, data: _*)
   }
+
+  test("special floating point values") {
+    import org.scalatest.exceptions.TestFailedException
+
+    // Spark treats -0.0 as 0.0
+    intercept[TestFailedException] {
+      checkDataset(Seq(-0.0d).toDS(), -0.0d)
+    }
+    intercept[TestFailedException] {
+      checkDataset(Seq(-0.0f).toDS(), -0.0f)
+    }
+    intercept[TestFailedException] {
+      checkDataset(Seq(Tuple1(-0.0)).toDS(), Tuple1(-0.0))
+    }
+
+    val floats = Seq[Float](-0.0f, 0.0f, Float.NaN).toDS()
+    checkDataset(floats, 0.0f, 0.0f, Float.NaN)
+
+    val doubles = Seq[Double](-0.0d, 0.0d, Double.NaN).toDS()
+    checkDataset(doubles, 0.0, 0.0, Double.NaN)
+
+    checkDataset(Seq(Tuple1(Float.NaN)).toDS(), Tuple1(Float.NaN))
+    checkDataset(Seq(Tuple1(-0.0f)).toDS(), Tuple1(0.0f))
+    checkDataset(Seq(Tuple1(Double.NaN)).toDS(), Tuple1(Double.NaN))
+    checkDataset(Seq(Tuple1(-0.0)).toDS(), Tuple1(0.0))
+
+    val complex = Map(Array(Seq(Tuple1(Double.NaN))) -> Map(Tuple2(Float.NaN, null)))
+    checkDataset(Seq(complex).toDS(), complex)
+  }
 }

sql/core/src/test/scala/org/apache/spark/sql/QueryTest.scala

@@ -132,6 +132,13 @@ abstract class QueryTest extends PlanTest {
       a.length == b.length && a.zip(b).forall { case (l, r) => compare(l, r)}
     case (a: Iterable[_], b: Iterable[_]) =>
       a.size == b.size && a.zip(b).forall { case (l, r) => compare(l, r)}
+    case (a: Product, b: Product) =>
+      compare(a.productIterator.toSeq, b.productIterator.toSeq)
+    // 0.0 == -0.0, turn float/double to binary before comparison, to distinguish 0.0 and -0.0.
+    case (a: Double, b: Double) =>
+      java.lang.Double.doubleToRawLongBits(a) == java.lang.Double.doubleToRawLongBits(b)
+    case (a: Float, b: Float) =>
+      java.lang.Float.floatToRawIntBits(a) == java.lang.Float.floatToRawIntBits(b)
     case (a, b) => a == b
   }