I've been using the built in EPiServer search data source to search through the pages and uploaded files on a project.  The resulting set seems to be fine, but all files always seem to come after pages.

The code used to create the searchdatasource:

            var searchDataSource = new SearchDataSource
                                       {
                                           SearchQuery = updatedSearchString,
                                           PageLink = this.SearchRoot,
                                           SearchFiles = true,
                                           SearchLocations = "~/Global/,~/PageFiles/",
                                           LanguageBranches = this.Languages,
                                           OnlyWholeWords = true
                                       };

SearchString returns the text from the search textbox.

The only modification being done with linq before binding(albeit quite a bit in the one modification):

var results = searchDataSource.Select(new DataSourceSelectArguments()).OfType<PageData>().Where(p => p["ExcludeFromSearchResults"] == null && !p.PageName.ToUpperInvariant().EndsWith(".XML")).Skip((pager.CurrentPageNumber * pager.PageWeight) - pager.PageWeight).Take(pager.PageWeight);

The purpose there is to ensure that anything with the ExcludeFromSearchResults page property enabled is ignored, and ignore any XML files.  Doing this after the query already has results is ugly, but I wanted to make sure files and pages would search as we expected before putting more time into optimizing the query.

Pager is a custom paging control we wrap around the repeater, PageWeight is the number of results being displayed per page.

Every set of search results seems to be ordered EPiServer Pages, then files.  The pages and files both appear to be properly sorted by relevance in their own right, but the list is always pages followed by files. 

Is there a way to get the two sets of results to be intermixed(when appropriate)?