[SPARK-23523][SQL][FOLLOWUP] Minor refactor of OptimizeMetadataOnlyQuery

## What changes were proposed in this pull request?

Inside `OptimizeMetadataOnlyQuery.getPartitionAttrs`, avoid using `zip` to generate attribute map.
Also include other minor update of comments and format.

## How was this patch tested?

Existing test cases.

Author: Xingbo Jiang <xingbo.jiang@databricks.com>

Closes #20693 from jiangxb1987/SPARK-23523.

sql/core/src/main/scala/org/apache/spark/sql/execution/OptimizeMetadataOnlyQuery.scala

@@ -83,7 +83,7 @@ case class OptimizeMetadataOnlyQuery(catalog: SessionCatalog) extends Rule[Logic
   private def getPartitionAttrs(
       partitionColumnNames: Seq[String],
       relation: LogicalPlan): Seq[Attribute] = {
-    val attrMap = relation.output.map(_.name.toLowerCase(Locale.ROOT)).zip(relation.output).toMap
+    val attrMap = relation.output.map(a => a.name.toLowerCase(Locale.ROOT) -> a).toMap
     partitionColumnNames.map { colName =>
       attrMap.getOrElse(colName.toLowerCase(Locale.ROOT),
         throw new AnalysisException(s"Unable to find the column `$colName` " +

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/HadoopFsRelation.scala

@@ -67,9 +67,9 @@ case class HadoopFsRelation(
     }
   }
 
-  // When data schema and partition schema have the overlapped columns, the output
-  // schema respects the order of data schema for the overlapped columns, but respect
-  // the data types of partition schema
+  // When data and partition schemas have overlapping columns, the output
+  // schema respects the order of the data schema for the overlapping columns, and it
+  // respects the data types of the partition schema.
   val schema: StructType = {
     StructType(dataSchema.map(f => overlappedPartCols.getOrElse(getColName(f), f)) ++
       partitionSchema.filterNot(f => overlappedPartCols.contains(getColName(f))))