[SPARK-22862] Docs on lazy elimination of columns missing from an encoder

This behavior has confused some users, so lets clarify it.

Author: Michael Armbrust <[email protected]>

Closes #20048 from marmbrus/datasetAsDocs.

Author: marmbrus

sql/core/src/main/scala/org/apache/spark/sql/Dataset.scala

@@ -401,6 +401,10 @@ class Dataset[T] private[sql](
    * If the schema of the Dataset does not match the desired `U` type, you can use `select`
    * along with `alias` or `as` to rearrange or rename as required.
    *
+   * Note that `as[]` only changes the view of the data that is passed into typed operations,
+   * such as `map()`, and does not eagerly project away any columns that are not present in
+   * the specified class.
+   *
    * @group basic
    * @since 1.6.0
    */