Preset Feature Feedback
What is the problem or goal you're trying to solve or accomplish?
When the Interactive Pivot Table groups data hierarchically (e.g., Channel > Region), derived ratio metrics are rolled up to parent levels by applying a single client-side aggregation (Sum, Average, Min, etc.) to the already-computed child values. For non-additive metrics this produces mathematically incorrect subtotals and totals.
Example: Average Handle Time is defined as total talk time ÷ total calls. With Region A = 100 min / 10 calls (AHT 10) and Region B = 100 min / 100 calls (AHT 1), the correct collapsed Channel-level AHT is 200 ÷ 110 = 1.82. The pivot table instead returns 11 (Sum) or 5.5 (Average) — a ratio cannot be recovered by summing or averaging child ratios. The parent-level value must be recalculated from the aggregated numerator and denominator.
How are you solving it currently?
There is no reliable workaround within the Interactive Pivot Table. The Value Aggregation menu operates on cell values AG Grid has already received and has no knowledge of the SQL metric definition. Customers fall back to either (a) a flat Table chart with custom expression metrics defined per grouping level — ergonomically poor and error-prone when upstream dimension values change, or (b) building separate charts per level.
What is your recommended solution?
Allow the pivot table to recompute derived/ratio metrics from their definitions at each aggregation level — by carrying the underlying numerator and denominator components and re-evaluating the ratio per group, rather than aggregating leaf-level results. A useful intermediate step would be per-metric (per-column) aggregation control, so additive measures and ratio measures are not forced to share a single aggregation function. Upstream references: apache/superset discussions #18548 and #38036.
