🤖 This PR was created by Repo Assist, an automated AI assistant.

Summary

Adds Frame.meltBy, a new frame-transformation function that keeps specified identity columns as-is and melts the remaining columns into Column/Value pairs. This is analogous to pandas DataFrame.melt(id_vars=...) and R's data.table::melt.

Closes #554.

Usage

let df =
    frame [ "ticker" =?> series [ 0 => "AAPL"; 1 => "MSFT" ]
            "price"  =?> series [ 0 => 189.5; 1 => 415.2 ]
            "volume" =?> series [ 0 => 1_200_000.0; 1 => 800_000.0 ] ]

let melted = df |> Frame.meltBy ["ticker"]
// Result:
// ticker | Column | Value
// -------+--------+----------
// AAPL   | price  | 189.5
// AAPL   | volume | 1200000.0
// MSFT   | price  | 415.2
// MSFT   | volume |  800000.0

Implementation

• Frame.meltBy (idCols: seq<'C>) (frame: Frame<'R, 'C>) in FrameModule.fs
• Accepts any seq<'C> of identity column keys (uses HashSet for O(1) lookup)
• Output frame: integer row index, string column keys ([id cols...] + "Column" + "Value")
• Preserves the original element type of each identity column via createTypedVector
• Infers the common supertype for the Value column (same strategy as Frame.melt)

Tests

4 new tests in Frame.fs:

• Row count is correct (rows × value_cols)
• Output column names include id columns + Column + Value
• Id column values are repeated correctly
• Id column element types are preserved

Test Status

✅ Build succeeded (0 errors, 4 warnings — all pre-existing)
✅ All 468 tests pass (including 4 new meltBy tests)

Generated by Repo Assist · ◷

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/repo-assist.md@30f2254f2a7a944da1224df45d181a3f8faefd0d