[SPARK-39159][SQL] Add new Dataset API for Offset

### What changes were proposed in this pull request?
Currently, Spark added `Offset` operator.
This PR try to add `offset` API into `Dataset`.

### Why are the changes needed?
`offset` API is very useful and construct test case more easily.

### Does this PR introduce _any_ user-facing change?
'No'.
New feature.

### How was this patch tested?
New tests.

Closes apache#36519 from beliefer/SPARK-39159.

Authored-by: Jiaan Geng <[email protected]>
Signed-off-by: Wenchen Fan <[email protected]>

Author: beliefer

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/CheckAnalysis.scala

@@ -404,13 +404,6 @@ trait CheckAnalysis extends PredicateHelper with LookupCatalog {
 
           case Offset(offsetExpr, _) => checkLimitLikeClause("offset", offsetExpr)
 
-          case o if !o.isInstanceOf[GlobalLimit] && !o.isInstanceOf[LocalLimit]
-            && o.children.exists(_.isInstanceOf[Offset]) =>
-            failAnalysis(
-              s"""
-                 |The OFFSET clause is allowed in the LIMIT clause or be the outermost node,
-                 |but the OFFSET clause found in: ${o.nodeName}.""".stripMargin.replace("\n", " "))
-
           case Tail(limitExpr, _) => checkLimitLikeClause("tail", limitExpr)
 
           case _: Union | _: SetOperation if operator.children.length > 1 =>

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/analysis/AnalysisErrorSuite.scala

@@ -541,13 +541,6 @@ class AnalysisErrorSuite extends AnalysisTest {
     "The offset expression must be equal to or greater than 0, but got -1" :: Nil
   )
 
-  errorTest(
-    "OFFSET clause in other node",
-    testRelation2.offset(Literal(10, IntegerType)).where('b > 1),
-    "The OFFSET clause is allowed in the LIMIT clause or be the outermost node," +
-      " but the OFFSET clause found in: Filter." :: Nil
-  )
-
   errorTest(
     "the sum of num_rows in limit clause and num_rows in offset clause less than Int.MaxValue",
     testRelation.offset(Literal(2000000000, IntegerType)).limit(Literal(1000000000, IntegerType)),

sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala

@@ -1955,6 +1955,16 @@ class Dataset[T] private[sql](
   def limitRange(start: Int, end: Int): Dataset[T] = withTypedPlan {
     Limit(Literal(end - start), Offset(Literal(start), logicalPlan))
   }
+  /**
+   * Returns a new Dataset by skipping the first `m` rows.
+   *
+   * @group typedrel
+   * @since 3.4.0
+   */
+  def offset(n: Int): Dataset[T] = withTypedPlan {
+    Offset(Literal(n), logicalPlan)
+  }
+
   /**
    * Returns a new Dataset containing union of rows in this Dataset and another Dataset.
    *

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

@@ -603,6 +603,30 @@ class DataFrameSuite extends QueryTest
     )
   }
 
+  test("offset") {
+    checkAnswer(
+      testData.offset(90),
+      testData.collect().drop(90).toSeq)
+
+    checkAnswer(
+      arrayData.toDF().offset(99),
+      arrayData.collect().drop(99).map(r => Row.fromSeq(r.productIterator.toSeq)))
+
+    checkAnswer(
+      mapData.toDF().offset(99),
+      mapData.collect().drop(99).map(r => Row.fromSeq(r.productIterator.toSeq)))
+  }
+
+  test("limit with offset") {
+    checkAnswer(
+      testData.limit(10).offset(5),
+      testData.take(10).drop(5).toSeq)
+
+    checkAnswer(
+      testData.offset(5).limit(10),
+      testData.take(15).drop(5).toSeq)
+  }
+
   test("udf") {
     val foo = udf((a: Int, b: String) => a.toString + b)