[SPARK-47309][SQL][XML] Fix schema inference issues in XML

### What changes were proposed in this pull request?

This PR fixes XML schema inference issues:

1. when there's an empty tag

2. when merging schema for NullType

### Why are the changes needed?

Fix a bug

### Does this PR introduce _any_ user-facing change?

Yes

### How was this patch tested?

Unit tests. There's a follow-up [PR](apache#45411) that introduces comprehensive tests for schema inference.

### Was this patch authored or co-authored using generative AI tooling?

No

Closes apache#45426 from shujingyang-db/fix-xml-schema-inference.

Authored-by: Shujing Yang <shujing.yang@databricks.com>
Signed-off-by: Hyukjin Kwon <gurwls223@apache.org>

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/xml/XmlInferSchema.scala

@@ -195,7 +195,9 @@ class XmlInferSchema(options: XmlOptions, caseSensitive: Boolean)
 
   private def inferField(parser: XMLEventReader): DataType = {
     parser.peek match {
-      case _: EndElement => NullType
+      case _: EndElement =>
+        parser.nextEvent()
+        NullType
       case _: StartElement => inferObject(parser)
       case _: Characters =>
         val structType = inferObject(parser).asInstanceOf[StructType]
@@ -450,7 +452,7 @@ class XmlInferSchema(options: XmlOptions, caseSensitive: Boolean)
     oldTypeOpt match {
       // If the field name already exists,
       // merge the type and infer the combined field as an array type if necessary
-      case Some(oldType) if !oldType.isInstanceOf[ArrayType] && !newType.isInstanceOf[NullType] =>
+      case Some(oldType) if !oldType.isInstanceOf[ArrayType] =>
         ArrayType(compatibleType(caseSensitive, options.valueTag)(oldType, newType))
       case Some(oldType) =>
         compatibleType(caseSensitive, options.valueTag)(oldType, newType)