Create Dictionary with LINQ and avoid "item with the same key has already been added" error

Question

I want to find a key in a dictionary and replace the value if it is found or add the key/value if it is not.

Code:

public class MyObject
{

    public string UniqueKey { get; set; }
    public string Field1 { get; set; }
    public string Field2 { get; set; }

}

LINQ Solution (throws An item with the same key has already been added.):

Dictionary<string, MyObject> objectDict = csvEntries.ToDictionary(csvEntry => csvEntry.ToMyObject().UniqueKey, csvEntry => csvEntry.ToMyObject());

ForEach solution (works):

Dictionary<string, MyObject> objectDict = new Dictionary<string, MyObject>();
foreach (CSVEntry csvEntry in csvEntries)
{

    MyObject obj = csvEntry.ToMyObject();

    if (objectDict.ContainsKey(obj.UniqueKey))
    {
        objectDict[obj.UniqueKey] = obj;
    }
    else {
        objectDict.Add(obj.UniqueKey, obj);
    }

}

I really liked the LINQ solution but as it stands, it throws the above error. Is there any nice way of avoiding the error and using LINQ?

Answer 1

You can use GroupBy to create unique keys:

Dictionary<string, MyObject> objectDict = csvEntries
    .Select(csvEntry => csvEntry.ToMyObject())
    .GroupBy(x => x.UniqueKey)
    .ToDictionary(grp => grp.Key, grp => grp.First());

However, instead of grp.First() you could create a collection with ToList or ToArray. On that way you don't take an arbitary object in case of duplicate keys.

Another option is to use a Lookup<TKey, TValue> which allows duplicate keys and even non-existing keys, you get an empty sequence in that case.

var uniqueKeyLookup = csvEntries
    .Select(csvEntry => csvEntry.ToMyObject())
    .ToLookup(x => x.UniqueKey);
IEnumerable<MyObject> objectsFor1234 = uniqueKeyLookup["1234"]; // empty if it doesn't exist

Answer 2

Building on Tim's answer, here's an extension method you can use so you don't need to duplicate the implementation throughout your project:

public static class DictionaryExtensions
{
    public static Dictionary<TKey, TValue> ToDictionaryWithDupSelector<TKey, TValue>(
        this IEnumerable<TValue> enumerable,
        Func<TValue, TKey> groupBy, Func<IEnumerable<TValue>, TValue> selector = null) {

        if (selector == null)
            selector = new Func<IEnumerable<TValue>, TValue>(grp => grp.First());

        return enumerable
            .GroupBy(e => groupBy(e))
            .ToDictionary(grp => grp.Key, grp => selector(grp));
    }
}

By default it will choose the first element when there are duplicates, but I have provided an optional parameter where you can specify an alternate selector. Example call to the extension method:

var objList = new List<string[]> {
    new string[2] {"1", "first"},
    new string[2] {"1", "last"},
    new string[2] {"2", "you"},
};
var asDict = objList.ToDictionary(
    arr => arr[0],
    grp => grp.Last()
);