[SPARK-24033][SQL] Fix Mismatched of Window Frame specifiedwindowframe(RowFrame, -1, -1)

## What changes were proposed in this pull request?

When the OffsetWindowFunction's frame is `UnaryMinus(Literal(1))` but the specified window frame has been simplified to `Literal(-1)` by some optimizer rules e.g., `ConstantFolding`. Thus, they do not match and cause the following error:
```
org.apache.spark.sql.AnalysisException: Window Frame specifiedwindowframe(RowFrame, -1, -1) must match the required frame specifiedwindowframe(RowFrame, -1, -1);
at org.apache.spark.sql.catalyst.analysis.CheckAnalysis$class.failAnalysis(CheckAnalysis.scala:41)
at org.apache.spark.sql.catalyst.analysis.Analyzer.failAnalysis(Analyzer.scala:91)
at
```
## How was this patch tested?
Added a test

Author: gatorsmile <[email protected]>

Closes apache#21115 from gatorsmile/fixLag.

Author: gatorsmile

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/windowExpressions.scala

@@ -342,7 +342,10 @@ abstract class OffsetWindowFunction
   override lazy val frame: WindowFrame = {
     val boundary = direction match {
       case Ascending => offset
-      case Descending => UnaryMinus(offset)
+      case Descending => UnaryMinus(offset) match {
+          case e: Expression if e.foldable => Literal.create(e.eval(EmptyRow), e.dataType)
+          case o => o
+      }
     }
     SpecifiedWindowFrame(RowFrame, boundary, boundary)
   }

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

@@ -402,4 +402,18 @@ class DataFrameWindowFramesSuite extends QueryTest with SharedSQLContext {
         Row(7, 3000) :: Row(8, 3000) :: Row(9, 5500) ::
         Row(10, 6000) :: Nil)
   }
+
+  test("SPARK-24033: Analysis Failure of OffsetWindowFunction") {
+    val ds = Seq((1, 1), (1, 2), (1, 3), (2, 1), (2, 2)).toDF("n", "i")
+    val res =
+      Row(1, 1, null) :: Row (1, 2, 1) :: Row(1, 3, 2) :: Row(2, 1, null) :: Row(2, 2, 1) :: Nil
+    checkAnswer(
+      ds.withColumn("m",
+        lead("i", -1).over(Window.partitionBy("n").orderBy("i").rowsBetween(-1, -1))),
+      res)
+    checkAnswer(
+      ds.withColumn("m",
+        lag("i", 1).over(Window.partitionBy("n").orderBy("i").rowsBetween(-1, -1))),
+      res)
+  }
 }