[SPARK-48808][SQL] Fix NPE when connecting thriftserver through Hive …

…1.2.1 and the result schema is empty

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

Lower Hive JDBC/Thrift Client doesn't test whether columns are null or not before handling it. In this PR, we set an empty list to it if the result schema is empty

### Why are the changes needed?

bugfix for lower hive clients

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

no

### How was this patch tested?

tested offline by wangyum

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

no

Closes apache#47213 from yaooqinn/SPARK-48808.

Lead-authored-by: Kent Yao <yao@apache.org>
Co-authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Kent Yao <yao@apache.org>

sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/RowSetUtils.scala

@@ -76,11 +76,12 @@ object RowSetUtils {
     val tRowSet = new TRowSet(startRowOffSet, new java.util.ArrayList[TRow](rowSize))
     var i = 0
     val columnSize = schema.length
+    val columns = new java.util.ArrayList[TColumn](columnSize)
     while (i < columnSize) {
-      val tColumn = toTColumn(rows, i, schema(i), timeFormatters, binaryFormatter)
-      tRowSet.addToColumns(tColumn)
+      columns.add(i, toTColumn(rows, i, schema(i), timeFormatters, binaryFormatter))
       i += 1
     }
+    tRowSet.setColumns(columns)
     tRowSet
   }