在DocumentDB LINQ查询中添加OrderBy子句会破坏它（C＃）

Question

I am using the Azure Documents SDK in C# in .net core 1.0.1 with LINQ to abstract a simple Fetch method that takes lambdas as parameters. It works fine EXCEPT when I add an ORDER BY clause – although I have no idea why. Here's the method:

public async Task<IEnumerable<T>> Fetch<R>(Expression<Func<T, bool>> predicate, Expression<Func<T, R>> orderby, int count)
{
    var results = new List<T>();
    var collectionLink = UriFactory.CreateDocumentCollectionUri(_config.Database, _collectionId);

    // This works
    //var document = _client.CreateDocumentQuery<T>(collectionLink).Where(predicate).Take(count).AsDocumentQuery();

    // This ** DOESN'T ** work
    var document = _client.CreateDocumentQuery<T>(collectionLink).Where(predicate).OrderBy(orderby).Take(count).AsDocumentQuery();

    var feedResponse = await document.ExecuteNextAsync<T>();
    results.AddRange(feedResponse.AsEnumerable<T>());
    return results;
}

With the addition of the .OrderBy() method, I get 0 results returned. With its removal, I get the expected 10 (albeit in insertion order). In looking at the SQL being generated by the .AsDocumentQuery() method, I can't see anything obviously wrong, so I'm at a bit of a loss as to why. For completeness, here's an example of its' usage...

var result = data.Fetch(f => f.Firstname.Contains("i"), x => x.Firstname, 10);

... and the query that it produces:

SELECT TOP 10 * FROM root WHERE CONTAINS(root["Firstname"], "i") ORDER BY root["Firstname"] ASC

Is this an SDK bug or am I missing something obvious?

Thanks!

Answer 1

Whenever I see strange behavior like this it's usually because you have a hash index (the default) as opposed to a range index on that field. Check the indexing policy for the collection.