Why C# doesn't implement indexed properties?

Question

I know, I know... Eric Lippert's answer to this kind of question is usually something like "because it wasn't worth the cost of designing, implementing, testing and documenting it".

But still, I'd like a better explanation... I was reading this blog post about new C# 4 features, and in the section about COM Interop, the following part caught my attention :

By the way, this code uses one more new feature: indexed properties (take a closer look at those square brackets after Range.) But this feature is available only for COM interop; you cannot create your own indexed properties in C# 4.0.

OK, but why ? I already knew and regretted that it wasn't possible to create indexed properties in C#, but this sentence made me think again about it. I can see several good reasons to implement it :

• the CLR supports it (for instance, PropertyInfo.GetValue has an index parameter), so it's a pity we can't take advantage of it in C#
• it is supported for COM interop, as shown in the article (using dynamic dispatch)
• it is implemented in VB.NET
• it is already possible to create indexers, i.e. to apply an index to the object itself, so it would probably be no big deal to extend the idea to properties, keeping the same syntax and just replacing this with a property name

It would allow to write that kind of things :

public class Foo
{
    private string[] _values = new string[3];
    public string Values[int index]
    {
        get { return _values[index]; }
        set { _values[index] = value; }
    }
}

Currently the only workaround that I know is to create an inner class (ValuesCollection for instance) that implements an indexer, and change the Values property so that it returns an instance of that inner class.

This is very easy to do, but annoying... So perhaps the compiler could do it for us ! An option would be to generate an inner class that implements the indexer, and expose it through a public generic interface :

// interface defined in the namespace System
public interface IIndexer<TIndex, TValue>
{
    TValue this[TIndex index]  { get; set; }
}

public class Foo
{
    private string[] _values = new string[3];

    private class <>c__DisplayClass1 : IIndexer<int, string>
    {
        private Foo _foo;
        public <>c__DisplayClass1(Foo foo)
        {
            _foo = foo;
        }

        public string this[int index]
        {
            get { return _foo._values[index]; }
            set { _foo._values[index] = value; }
        }
    }

    private IIndexer<int, string> <>f__valuesIndexer;
    public IIndexer<int, string> Values
    {
        get
        {
            if (<>f__valuesIndexer == null)
                <>f__valuesIndexer = new <>c__DisplayClass1(this);
            return <>f__valuesIndexer;
        }
    }
}

But of course, in that case the property would actually return a IIndexer<int, string>, and wouldn't really be an indexed property... It would be better to generate a real CLR indexed property.

What do you think ? Would you like to see this feature in C# ? If not, why ?

Answer 1

Here's how we designed C# 4.

First we made a list of every possible feature we could think of adding to the language.

Then we bucketed the features into "this is bad, we must never do it", "this is awesome, we have to do it", and "this is good but let's not do it this time".

Then we looked at how much budget we had to design, implement, test, document, ship and maintain the "gotta have" features and discovered that we were 100% over budget.

So we moved a bunch of stuff from the "gotta have" bucket to the "nice to have" bucket.

Indexed properties were never anywhere near the top of the "gotta have" list. They are very low on the "nice" list and flirting with the "bad idea" list.

Every minute we spend designing, implementing, testing, documenting or maintaining nice feature X is a minute we can't spend on awesome features A, B, C, D, E, F and G. We have to ruthlessly prioritize so that we only do the best possible features. Indexed properties would be nice, but nice isn't anywhere even close to good enough to actually get implemented.

Answer 2

A C# indexer is an indexed property. It is named Item by default (and you can refer to it as such from e.g. VB), and you can change it with IndexerNameAttribute if you want.

I'm not sure why, specifically, it was designed that way, but it does seem to be an intentional limitation. It is, however, consistent with Framework Design Guidelines, which do recommend the approach of a non-indexed property returning an indexable object for member collections. I.e. "being indexable" is a trait of a type; if it's indexable in more than one way, then it really should be split into several types.

Answer 3

Because you can already do it kind of, and it's forced you to think in OO aspects, adding indexed properties would just add more noise to the language. And just another way to do another thing.

class Foo
{
    public Values Values { ... }
}

class Values
{
    public string this[int index] { ... }    
}

foo.Values[0]

I personally would prefer to see only a single way of doing something, rather than 10 ways. But of course this is a subjective opinion.

Answer 4

I used to favor the idea of indexed properties but then realized it would add horrible ambiguity and actually disincentivize functionality. Indexed properties would mean you don't have a child collection instance. That's both good and bad. It's less trouble to implement and you don't need a reference back to the enclosing owner class. But it also means you can't pass that child collection to anything; you'd likely have to enumerate every single time. Nor can you do a foreach on it. Worst of all, you can't tell from looking at an indexed property whether it's that or a collection property.

The idea is rational but it just leads to inflexibility and abrupt awkwardness.

Answer 5

I find the lack of indexed properties very frustrating when trying to write clean, concise code. An indexed property has a very different connotation than providing a class reference that's indexed or providing individual methods. I find it a bit disturbing that providing access to an internal object that implements an indexed property is even considered acceptable since that often breaks one of the key components of object orientation: encapsulation.

I run into this problem often enough, but I just encountered it again today so I'll provide a real world code example. The interface and class being written stores application configuration which is a collection of loosely related information. I needed to add named script fragments and using the unnamed class indexer would have implied a very wrong context since script fragments are only part of the configuration.

If indexed properties were available in C# I could have implemented the below code (syntax is this[key] changed to PropertyName[key]).

public interface IConfig
{
    // Other configuration properties removed for examp[le

    /// <summary>
    /// Script fragments
    /// </summary>
    string Scripts[string name] { get; set; }
}

/// <summary>
/// Class to handle loading and saving the application's configuration.
/// </summary>
internal class Config : IConfig, IXmlConfig
{
  #region Application Configuraiton Settings

    // Other configuration properties removed for examp[le

    /// <summary>
    /// Script fragments
    /// </summary>
    public string Scripts[string name]
    {
        get
        {
            if (!string.IsNullOrWhiteSpace(name))
            {
                string script;
                if (_scripts.TryGetValue(name.Trim().ToLower(), out script))
                    return script;
            }
            return string.Empty;
        }
        set
        {
            if (!string.IsNullOrWhiteSpace(name))
            {
                _scripts[name.Trim().ToLower()] = value;
                OnAppConfigChanged();
            }
        }
    }
    private readonly Dictionary<string, string> _scripts = new Dictionary<string, string>();

  #endregion

    /// <summary>
    /// Clears configuration settings, but does not clear internal configuration meta-data.
    /// </summary>
    private void ClearConfig()
    {
        // Other properties removed for example
        _scripts.Clear();
    }

  #region IXmlConfig

    void IXmlConfig.XmlSaveTo(int configVersion, XElement appElement)
    {
        Debug.Assert(configVersion == 2);
        Debug.Assert(appElement != null);

        // Saving of other properties removed for example

        if (_scripts.Count > 0)
        {
            var scripts = new XElement("Scripts");
            foreach (var kvp in _scripts)
            {
                var scriptElement = new XElement(kvp.Key, kvp.Value);
                scripts.Add(scriptElement);
            }
            appElement.Add(scripts);
        }
    }

    void IXmlConfig.XmlLoadFrom(int configVersion, XElement appElement)
    {
        // Implementation simplified for example

        Debug.Assert(appElement != null);
        ClearConfig();
        if (configVersion == 2)
        {
            // Loading of other configuration properites removed for example

            var scripts = appElement.Element("Scripts");
            if (scripts != null)
                foreach (var script in scripts.Elements())
                    _scripts[script.Name.ToString()] = script.Value;
        }
        else
            throw new ApplicaitonException("Unknown configuration file version " + configVersion);
    }

  #endregion
}

Unfortunately indexed properties are not implemented so I implemented a class to store them and provided access to that. This is an undesirable implementation because the purpose of the configuration class in this domain model is to encapsulate all the details. Clients of this class will be accessing specific script fragments by name and have no reason to count or enumerate over them.

I could have implemented this as:

public string ScriptGet(string name)
public void ScriptSet(string name, string value)

Which I probably should have, but this is a useful illustration of why using indexed classes as a replacement for this missing feature is often not a reasonable substitute.

To implement similar capability as an indexed property I had to write the below code which you'll notice is considerably longer, more complex and thus harder to read, understand and maintain.

public interface IConfig
{
    // Other configuration properties removed for examp[le

    /// <summary>
    /// Script fragments
    /// </summary>
    ScriptsCollection Scripts { get; }
}

/// <summary>
/// Class to handle loading and saving the application's configuration.
/// </summary>
internal class Config : IConfig, IXmlConfig
{
    public Config()
    {
        _scripts = new ScriptsCollection();
        _scripts.ScriptChanged += ScriptChanged;
    }

  #region Application Configuraiton Settings

    // Other configuration properties removed for examp[le

    /// <summary>
    /// Script fragments
    /// </summary>
    public ScriptsCollection Scripts
    { get { return _scripts; } }
    private readonly ScriptsCollection _scripts;

    private void ScriptChanged(object sender, ScriptChangedEventArgs e)
    {
        OnAppConfigChanged();
    }

  #endregion

    /// <summary>
    /// Clears configuration settings, but does not clear internal configuration meta-data.
    /// </summary>
    private void ClearConfig()
    {
        // Other properties removed for example
        _scripts.Clear();
    }

  #region IXmlConfig

    void IXmlConfig.XmlSaveTo(int configVersion, XElement appElement)
    {
        Debug.Assert(configVersion == 2);
        Debug.Assert(appElement != null);

        // Saving of other properties removed for example

        if (_scripts.Count > 0)
        {
            var scripts = new XElement("Scripts");
            foreach (var kvp in _scripts)
            {
                var scriptElement = new XElement(kvp.Key, kvp.Value);
                scripts.Add(scriptElement);
            }
            appElement.Add(scripts);
        }
    }

    void IXmlConfig.XmlLoadFrom(int configVersion, XElement appElement)
    {
        // Implementation simplified for example

        Debug.Assert(appElement != null);
        ClearConfig();
        if (configVersion == 2)
        {
            // Loading of other configuration properites removed for example

            var scripts = appElement.Element("Scripts");
            if (scripts != null)
                foreach (var script in scripts.Elements())
                    _scripts[script.Name.ToString()] = script.Value;
        }
        else
            throw new ApplicaitonException("Unknown configuration file version " + configVersion);
    }

  #endregion
}

public class ScriptsCollection : IEnumerable<KeyValuePair<string, string>>
{
    private readonly Dictionary<string, string> Scripts = new Dictionary<string, string>();

    public string this[string name]
    {
        get
        {
            if (!string.IsNullOrWhiteSpace(name))
            {
                string script;
                if (Scripts.TryGetValue(name.Trim().ToLower(), out script))
                    return script;
            }
            return string.Empty;
        }
        set
        {
            if (!string.IsNullOrWhiteSpace(name))
                Scripts[name.Trim().ToLower()] = value;
        }
    }

    public void Clear()
    {
        Scripts.Clear();
    }

    public int Count
    {
        get { return Scripts.Count; }
    }

    public event EventHandler<ScriptChangedEventArgs> ScriptChanged;

    protected void OnScriptChanged(string name)
    {
        if (ScriptChanged != null)
        {
            var script = this[name];
            ScriptChanged.Invoke(this, new ScriptChangedEventArgs(name, script));
        }
    }

  #region IEnumerable

    public IEnumerator<KeyValuePair<string, string>> GetEnumerator()
    {
        return Scripts.GetEnumerator();
    }

    IEnumerator IEnumerable.GetEnumerator()
    {
        return GetEnumerator();
    }

  #endregion
}

public class ScriptChangedEventArgs : EventArgs
{
    public string Name { get; set; }
    public string Script { get; set; }

    public ScriptChangedEventArgs(string name, string script)
    {
        Name = name;
        Script = script;
    }
}

Answer 6

Another workaround is listed at Easy creation of properties that support indexing in C#, that requires less work.

EDIT: I should also add that in response to the original question, that my if we can accomplish the desired syntax, with library support, then I think there needs to be a very strong case to add it directly to the language, in order to minimize language bloat.

Answer 7

There is a simple general solution using lambdas to proxy the indexing functionality

For read only indexing

public class RoIndexer<TIndex, TValue>
{
    private readonly Func<TIndex, TValue> _Fn;

    public RoIndexer(Func<TIndex, TValue> fn)
    {
        _Fn = fn;
    }

    public TValue this[TIndex i]
    {
        get
        {
            return _Fn(i);
        }
    }
}

For mutable indexing

public class RwIndexer<TIndex, TValue>
{
    private readonly Func<TIndex, TValue> _Getter;
    private readonly Action<TIndex, TValue> _Setter;

    public RwIndexer(Func<TIndex, TValue> getter, Action<TIndex, TValue> setter)
    {
        _Getter = getter;
        _Setter = setter;
    }

    public TValue this[TIndex i]
    {
        get
        {
            return _Getter(i);
        }
        set
        {
            _Setter(i, value);
        }
    }
}

and a factory

public static class Indexer
{
    public static RwIndexer<TIndex, TValue> Create<TIndex, TValue>(Func<TIndex, TValue> getter, Action<TIndex, TValue> setter)
    {
        return new RwIndexer<TIndex, TValue>(getter, setter);
    } 
    public static RoIndexer<TIndex, TValue> Create<TIndex, TValue>(Func<TIndex, TValue> getter)
    {
        return new RoIndexer<TIndex, TValue>(getter);
    } 
}

in my own code I use it like

public class MoineauFlankContours
{

    public MoineauFlankContour Rotor { get; private set; }

    public MoineauFlankContour Stator { get; private set; }

     public MoineauFlankContours()
    {
        _RoIndexer = Indexer.Create(( MoineauPartEnum p ) => 
            p == MoineauPartEnum.Rotor ? Rotor : Stator);
    }
    private RoIndexer<MoineauPartEnum, MoineauFlankContour> _RoIndexer;

    public RoIndexer<MoineauPartEnum, MoineauFlankContour> FlankFor
    {
        get
        {
            return _RoIndexer;
        }
    }

}

and with an instance of MoineauFlankContours I can do

MoineauFlankContour rotor = contours.FlankFor[MoineauPartEnum.Rotor];
MoineauFlankContour stator = contours.FlankFor[MoineauPartEnum.Stator];

Answer 8

Well I would say that they haven't added it because it wasn't worth the cost of designing, implementing, testing and documenting it.

Joking aside, its probably because the workarounds are simple and the feature never makes the time versus benefit cut. I wouldn't be surprised to see this appear as a change down the line though.

You also forgot to mention that an easier workaround is just make a regular method:

public void SetFoo(int index, Foo toSet) {...}
public Foo GetFoo(int index) {...}

Answer 9

Just found out myself too that you can use explicitly implemented interfaces to achieve this, as shown here: Named indexed property in C#? (see the second way shown in that reply)