Update the scroll example in the docs (#37394)

Update the scroll example ascii and Java docs, so it is more clear when to 
consume the scroll documents. Before this change the user could loose 
the first results if one uses copy & paste.

client/rest-high-level/src/test/java/org/elasticsearch/client/documentation/SearchDocumentationIT.java

@@ -715,12 +715,12 @@ public void onFailure(Exception e) {
             SearchHit[] searchHits = searchResponse.getHits().getHits();
 
             while (searchHits != null && searchHits.length > 0) { // <2>
-                SearchScrollRequest scrollRequest = new SearchScrollRequest(scrollId); // <3>
+                // <3>
+                SearchScrollRequest scrollRequest = new SearchScrollRequest(scrollId); // <4>
                 scrollRequest.scroll(scroll);
                 searchResponse = client.scroll(scrollRequest, RequestOptions.DEFAULT);
                 scrollId = searchResponse.getScrollId();
                 searchHits = searchResponse.getHits().getHits();
-                // <4>
             }
 
             ClearScrollRequest clearScrollRequest = new ClearScrollRequest(); // <5>

docs/java-rest/high-level/search/scroll.asciidoc

@@ -128,9 +128,9 @@ include-tagged::{doc-tests}/SearchDocumentationIT.java[search-scroll-example]
 <1> Initialize the search context by sending the initial `SearchRequest`
 <2> Retrieve all the search hits by calling the Search Scroll api in a loop
 until no documents are returned
-<3> Create a new `SearchScrollRequest` holding the last returned scroll
+<3> Process the returned search results
+<4> Create a new `SearchScrollRequest` holding the last returned scroll
 identifier and the scroll interval
-<4> Process the returned search results
 <5> Clear the scroll context once the scroll is completed
 
 [[java-rest-high-clear-scroll]]