Pivot tricks

Pivoting is the fastest way from a flat list of facts to an answer-shaped table. Every cell on this page is live and executable over the same tiny in-RAM cube — a flat Sale(Region, Product, Year, Amount) array — and opens with the prompt that would produce it. The pivot API lives in MeshWeaver.Layout.Pivot (ToPivotGrid + GroupRowsBy / GroupColumnsBy / Aggregate); the chart-side twin of the same slicing idea is in Charts at a glance.

Rows by X, columns by Y

Prompt: "Pivot sales: regions as rows, years as columns, sum of amount."

using MeshWeaver.Layout.Pivot;

record Sale(string Region, string Product, string Year, double Amount);

var sales = new[]
{
    new Sale("EMEA", "Espresso", "2024", 210), new Sale("EMEA", "Espresso", "2025", 260),
    new Sale("EMEA", "Grinder", "2024", 110),  new Sale("EMEA", "Grinder", "2025", 130),
    new Sale("Americas", "Espresso", "2024", 180), new Sale("Americas", "Espresso", "2025", 215),
    new Sale("Americas", "Grinder", "2024", 100),  new Sale("Americas", "Grinder", "2025", 120),
    new Sale("APAC", "Espresso", "2024", 90),  new Sale("APAC", "Espresso", "2025", 135),
    new Sale("APAC", "Grinder", "2024", 60),   new Sale("APAC", "Grinder", "2025", 70),
};

sales.ToPivotGrid(pivot => pivot
    .GroupRowsBy(s => s.Region)
    .GroupColumnsBy(s => s.Year)
    .Aggregate(s => s.Amount, agg => agg.WithFunction(AggregateFunction.Sum)))

Totals — close the table

Prompt: "Same pivot, but add row and column totals so I can sanity-check the sums."

using MeshWeaver.Layout.Pivot;

record Sale(string Region, string Product, string Year, double Amount);

var sales = new[]
{
    new Sale("EMEA", "Espresso", "2024", 210), new Sale("EMEA", "Espresso", "2025", 260),
    new Sale("EMEA", "Grinder", "2024", 110),  new Sale("EMEA", "Grinder", "2025", 130),
    new Sale("Americas", "Espresso", "2024", 180), new Sale("Americas", "Espresso", "2025", 215),
    new Sale("Americas", "Grinder", "2024", 100),  new Sale("Americas", "Grinder", "2025", 120),
    new Sale("APAC", "Espresso", "2024", 90),  new Sale("APAC", "Espresso", "2025", 135),
    new Sale("APAC", "Grinder", "2024", 60),   new Sale("APAC", "Grinder", "2025", 70),
};

sales.ToPivotGrid(pivot => pivot
    .GroupRowsBy(s => s.Region)
    .GroupColumnsBy(s => s.Year)
    .Aggregate(s => s.Amount, agg => agg.WithFunction(AggregateFunction.Sum))
    .WithRowTotals()
    .WithColumnTotals())

Two row dimensions work the same way — chain a second GroupRowsBy(s => s.Product) and the grid nests products under regions.

Top-N — cut the long tail

Top-N is a LINQ fold before the render: aggregate, order, take, and hand the finished rows to a DataGrid.

Prompt: "Give me the top 3 region-product combinations by total sales, as a table."

record Sale(string Region, string Product, string Year, double Amount);

var sales = new[]
{
    new Sale("EMEA", "Espresso", "2024", 210), new Sale("EMEA", "Espresso", "2025", 260),
    new Sale("EMEA", "Grinder", "2024", 110),  new Sale("EMEA", "Grinder", "2025", 130),
    new Sale("Americas", "Espresso", "2024", 180), new Sale("Americas", "Espresso", "2025", 215),
    new Sale("Americas", "Grinder", "2024", 100),  new Sale("Americas", "Grinder", "2025", 120),
    new Sale("APAC", "Espresso", "2024", 90),  new Sale("APAC", "Espresso", "2025", 135),
    new Sale("APAC", "Grinder", "2024", 60),   new Sale("APAC", "Grinder", "2025", 70),
};

record TopRow(string Region, string Product, double Total);

var top3 = sales
    .GroupBy(s => (s.Region, s.Product))
    .Select(g => new TopRow(g.Key.Region, g.Key.Product, g.Sum(s => s.Amount)))
    .OrderByDescending(r => r.Total)
    .Take(3)
    .ToArray();

Controls.DataGrid(top3)

YoY deltas — the comparison column

Year-over-year is a self-join on the year dimension: pivot the two years side by side in LINQ, compute the delta, render the finished comparison.

Prompt: "Compare 2025 vs 2024 sales per region and show the growth in percent."

record Sale(string Region, string Product, string Year, double Amount);

var sales = new[]
{
    new Sale("EMEA", "Espresso", "2024", 210), new Sale("EMEA", "Espresso", "2025", 260),
    new Sale("EMEA", "Grinder", "2024", 110),  new Sale("EMEA", "Grinder", "2025", 130),
    new Sale("Americas", "Espresso", "2024", 180), new Sale("Americas", "Espresso", "2025", 215),
    new Sale("Americas", "Grinder", "2024", 100),  new Sale("Americas", "Grinder", "2025", 120),
    new Sale("APAC", "Espresso", "2024", 90),  new Sale("APAC", "Espresso", "2025", 135),
    new Sale("APAC", "Grinder", "2024", 60),   new Sale("APAC", "Grinder", "2025", 70),
};

record YoyRow(string Region, double Y2024, double Y2025, string Growth);

var yoy = sales
    .GroupBy(s => s.Region)
    .Select(g =>
    {
        var y24 = g.Where(s => s.Year == "2024").Sum(s => s.Amount);
        var y25 = g.Where(s => s.Year == "2025").Sum(s => s.Amount);
        return new YoyRow(g.Key, y24, y25, $"{(y25 - y24) / y24:P1}");
    })
    .OrderByDescending(r => r.Y2025)
    .ToArray();

Controls.DataGrid(yoy)

The same fold feeds a chart directly — pipe yoy into the SliceBy(...).ToColumnChart(...) pipeline from Charts at a glance and the comparison becomes grouped columns.

Try these in the Agentic Engineering exercises.