[SPARK-49275][SQL] Fix return type nullness of the xpath expression

The `xpath` expression incorrectly marks its return type as array of non-null strings. However, it can actually return an array containing nulls. This can cause NPE in code generation, such as query `select coalesce(xpath(repeat('<a></a>', id), 'a')[0], '') from range(1, 2)`.

It avoids potential failures in queries that uses the `xpath` expression.

No.

A new unit test. It would fail without the change in the PR.

No.

Closes apache#47796 from chenhao-db/fix_xpath_nullness.

Authored-by: Chenhao Li <[email protected]>
Signed-off-by: Max Gekk <[email protected]>

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/xml/xpath.scala

@@ -239,13 +239,16 @@ case class XPathString(xml: Expression, path: Expression) extends XPathExtract {
     Examples:
       > SELECT _FUNC_('<a><b>b1</b><b>b2</b><b>b3</b><c>c1</c><c>c2</c></a>','a/b/text()');
        ["b1","b2","b3"]
+      > SELECT _FUNC_('<a><b>b1</b><b>b2</b><b>b3</b><c>c1</c><c>c2</c></a>','a/b');
+       [null,null,null]
   """,
   since = "2.0.0",
   group = "xml_funcs")
 // scalastyle:on line.size.limit
 case class XPathList(xml: Expression, path: Expression) extends XPathExtract {
   override def prettyName: String = "xpath"
-  override def dataType: DataType = ArrayType(StringType, containsNull = false)
+
+  override def dataType: DataType = ArrayType(StringType)
 
   override def nullSafeEval(xml: Any, path: Any): Any = {
     val nodeList = xpathUtil.evalNodeList(xml.asInstanceOf[UTF8String].toString, pathString)

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/expressions/xml/XPathExpressionSuite.scala

@@ -185,6 +185,11 @@ class XPathExpressionSuite extends SparkFunSuite with ExpressionEvalHelper {
     testExpr("<a><b class='bb'>b1</b><b>b2</b><b>b3</b><c class='bb'>c1</c><c>c2</c></a>",
       "a/*[@class='bb']/text()", Seq("b1", "c1"))
 
+    checkEvaluation(
+      Coalesce(Seq(
+          GetArrayItem(XPathList(Literal("<a></a>"), Literal("a")), Literal(0)),
+          Literal("nul"))), "nul")
+
     testNullAndErrorBehavior(testExpr)
   }