[HUDI-4296] Fix the bug that TestHoodieSparkSqlWriter.testSchemaEvolutionForTableType is flaky (#5973)

hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/hudi/BaseFileOnlyRelation.scala

@@ -166,7 +166,9 @@ class BaseFileOnlyRelation(sqlContext: SQLContext,
       DataSource.apply(
         sparkSession = sparkSession,
         paths = extraReadPaths,
-        userSpecifiedSchema = userSchema,
+        // Here we should specify the schema to the latest commit schema since
+        // the table schema evolution.
+        userSpecifiedSchema = userSchema.orElse(Some(tableStructSchema)),
         className = formatClassName,
         // Since we're reading the table as just collection of files we have to make sure
         // we only read the latest version of every Hudi's file-group, which might be compacted, clustered, etc.
@@ -175,8 +177,7 @@ class BaseFileOnlyRelation(sqlContext: SQLContext,
         // We rely on [[HoodieROTablePathFilter]], to do proper filtering to assure that
         options = optParams ++ Map(
           "mapreduce.input.pathFilter.class" -> classOf[HoodieROTablePathFilter].getName
-        ),
-        partitionColumns = partitionColumns
+        )
       )
         .resolveRelation()
         .asInstanceOf[HadoopFsRelation]