dotnet/machinelearning · error · NotImplementedException

nameof(TKey)

Error message

nameof(TKey)

What it means

Marks the not-implemented branch in StringDataFrameColumn's grouped-indices method (the typed GetGroupedIndices<TKey> path). Grouping is only implemented for TKey = string, because values are bucketed via a string multimap; any other key type falls into the NotImplementedException that names nameof(TKey) as the offending parameter.

Solutions

1. Use one of the supported TKey types (e.g. long, int, string, bool, DateTime per the API docs)
2. Pre-convert the key column to a supported primitive type before grouping
3. Implement custom grouping with LINQ GroupBy if the key type is unsupported
4. Check the source for the supported TKey set and match it

Example fix

// before
var grouped = column.GetGroupedIndices<MyKey>(...);
// after
var grouped = column.GetGroupedIndices<long>(key => key.GetHashCode());

if (!(typeof(TKey) == typeof(long) || typeof(TKey) == typeof(int) || typeof(TKey) == typeof(string) || typeof(TKey) == typeof(bool)))
    throw new NotSupportedException($"TKey {typeof(TKey)} not supported for grouping");

bool IsSupportedKeyType<TKey>() => typeof(TKey) == typeof(long) || typeof(TKey) == typeof(int) || typeof(TKey) == typeof(bool);

try { column.GetGroupedIndices<TKey>(...); } catch (NotImplementedException) { /* fall back to LINQ GroupBy */ }

When it happens

Trigger: Calling GetGroupedIndices<TKey> (or similar grouping API) with an unsupported key type such as a custom class, Guid, or short.

Common situations: Grouping by an unexpected key type inferred from generics; migrating code that assumed arbitrary key types are supported.

Related errors

• Cannot cast elements of column
• nameof(TKey)

AI-assisted analysis of dotnet/machinelearning@7b76e69cf9 (2026-09-11). Data as JSON: /api/errors/2c50eb6c856c5c4c. Report an issue: GitHub.

Appendix: source

Thrown at src/Microsoft.Data.Analysis/DataFrameColumns/StringDataFrameColumn.cs:438

                        if (containsKey)
                        {
                            values.Add(i);
                        }
                        else
                        {
                            multimap.Add(str, new List<long>() { i });
                        }
                    }
                    else
                    {
                        nullIndices.Add(i);
                    }
                }
                return multimap as Dictionary<TKey, ICollection<long>>;
            }
            else
            {
                throw new NotImplementedException(nameof(TKey));
            }
        }

        /// <summary>
        /// Returns a new column with <see langword="null" /> elements replaced by <paramref name="value"/>.
        /// </summary>
        /// <remarks>Tries to convert value to the column's DataType</remarks>
        /// <param name="value"></param>
        /// <param name="inPlace">Indicates if the operation should be performed in place</param>
        public StringDataFrameColumn FillNulls(string value, bool inPlace = false)
        {
            if (value == null)
                throw new ArgumentNullException(nameof(value));
            StringDataFrameColumn column = inPlace ? this : Clone();

            for (long i = 0; i < column.Length; i++)
            {
                if (column[i] == null)

View on GitHub (pinned to 7b76e69cf9)