With ExamineFileIndexer, I get duplicate results when:

1. Using a MultiIndexSearcher, and
2. The same term is found in two files

I do not get duplicate results when the same term is found in two document types.

Here's the element I added for the searcher:

<add name="MultiIndexSearcher" type="Examine.LuceneEngine.Providers.MultiIndexSearcher, Examine" analyzer="Lucene.Net.Analysis.Standard.StandardAnalyzer, Lucene.Net" indexSets="ExternalIndexSet,MediaIndexSet" enableLeadingWildcards="false" />

The ExternalIndexSet and the MediaIndexSet referred to are the out-of-box set and the media set inserted when I installed the ExamineFileIndexer package, respectively.

Here's code from my test page:

    @using Examine.SearchCriteria
@using Examine
@{
    Layout = null;

    string term = "control";

    // Get a searcher
    Examine.Providers.BaseSearchProvider searcher = ExamineManager.Instance.SearchProviderCollection["MultiIndexSearcher"];

    // Create search criteria instance
    ISearchCriteria searchCriteria = searcher.CreateSearchCriteria();

    // Pass the search criteria instance our search term as a raw query
    ISearchCriteria query = searchCriteria.RawQuery(term);

    // Search using the criteria in our query. Receive results.
    ISearchResults searchResults = searcher.Search(query);

}

<ul>
    @foreach (var searchResult in searchResults)
    {
        if (searchResult.Fields["nodeTypeAlias"].ToString().ToLower() == "file")
        {
            //we have found a physical file in the media folder structure
            //note that this is the line where we can capitalise on that Media File custom property we created - friendlyFileName
            var media = Umbraco.Media(searchResult.Id);
    <li>
        <a href="@media.Url" target="_blank">@searchResult.Fields["nodeName"]</a>
    </li>
    }
    else
    {
    //we are a content type other than a file
    <li>
        <a href="@Umbraco.NiceUrl(searchResult.Id)">@searchResult.Fields["nodeName"]</a>
    </li>
    }
    }
</ul>

In my database, there are two .pdf files and two document types with the word "control" in them.

When I use the MultiIndexSearcher, I get six results: one each for the doc types, and two each for the files (duplicates, in other words).

When I use the MediaIndexSearcher, I get two results -- one for each file. Also, if I remove the ExternalIndexSet from the MultiIndexSearcher, I get the proper number of results (one for each file).

I know I could, of course, query the indexes separately and combine the results. I'd rather figure out how to make the MultiIndexSearcher work, though. Any help would be appreciated!