Group

Calculates measures on input records, with optional groupings:

Example without groups

// this stage returns a single row with one column per measure
{
    measures: {
        "avg_sales": "avg(sales)",
        "total_sales": "sum(sales)"
    }
}

Optionally, groupBy may contain a list of expressions by which to group the calculations. This is useful for comparing results of mutually exclusive groupings. For overlapping groupings, use the Aggregate stage instead.

Example with groups

// this stage returns a row for every discovered combination of region and metro
// it returns four columns: region, metric, avg_sales, and total_sales
{
    groupBy: ["region", "metro"],
    measures: {
        "avg_sales": "avg(sales)",
        "total_sales": "sum(sales)"
    }
}

Each GroupBy entry can be an object, to include additional field meta. Measures too can be objects.

Example with Field Meta

// this calcultes the same as above, but returns additional Field Meta to help a visualizer
{
    groupBy: [
        { syntax: "region", label: "US Region" },
        { syntax: "metro", label: "Metropolitan Area" }
    ],
    measures: [
        { syntax: "avg(sales)", label: "Average sales", format: "$.2f" },
        { syntax: "sum(sales)", label: "Total sales", format: "$.2f" }
    ]
}