Add doc for returned value.

Author: yhuai

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala

@@ -651,6 +651,9 @@ class Analyzer(
      * the window expression as attribute references. So, the first returned value will be
      * `[Sum(_w0) OVER (PARTITION BY _w1 ORDER BY _w2)]` and the second returned value will be
      * [col1, col2 + col3 as _w0, col4 as _w1, col5 as _w2].
+     *
+     * @return (seq of expressions containing at lease one window expressions,
+     *          seq of non-window expressions)
      */
     private def extract(
         expressions: Seq[NamedExpression]): (Seq[NamedExpression], Seq[NamedExpression]) = {