@@ -0,0 +1,101 @@
+// Copyright (c) 2026 ObjectStack. Licensed under the Apache-2.0 license.
+
+/**
+ * #20727 — the four drill-through sidecars are set on the declared
+ * `AnalyticsResult`, with nothing in between.
+ *
+ * `AnalyticsResult` (`@objectstack/spec/contracts`) declares `dimensionFields`,
+ * `drillRawRows`, `drillRawTotals` and `drillRanges` (#20700). The service once
+ * set them through a module-private augmentation of that type; it now sets the
+ * declared members directly. That retirement moved no byte of any answer, and
+ * this file holds that line.
+ *
+ * The fixture is the one answer that carries all four at once: a grouping by
+ * an equality-drillable dimension (`stage`) beside a month-bucketed date
+ * dimension (`close_date`), with a `totals` selection (a per-stage subtotal
+ * and the grand total). The answer is read as `AnalyticsResult` with no cast,
+ * so each member read below compiles only because the contract declares it.
+ */
+
+import { describe, it, expect } from 'vitest';
+import { DatasetSchema } from '@objectstack/spec/ui';
+import type { ExecutionContext } from '@objectstack/spec/kernel';
+import type { AnalyticsResult } from '@objectstack/spec/contracts';
+import { AnalyticsService } from '../analytics-service.js';
+
+const CTX = { tenantId: 'org_A' } as ExecutionContext;
+
+const DATASET = DatasetSchema.parse({
+  name: 'pipe_drill',
+  label: 'Pipe',
+  object: 'opportunity',
+  include: [],
+  dimensions: [
+    { name: 'stage', field: 'stage', type: 'string' },
+    { name: 'close_date', field: 'close_date', type: 'date', dateGranularity: 'month' },
+  ],
+  measures: [{ name: 'cnt', aggregate: 'count' }],
+});
+
+/** The grouped field names of one `executeAggregate` call (a date bucket arrives as `{ field, dateGranularity }`). */
+function groupedFields(groupBy: unknown[] | undefined): string[] {
+  return (groupBy ?? []).map((g) => (typeof g === 'string' ? g : (g as { field: string }).field));
+}
+
+/**
+ * Granularity bucketing runs through the ObjectQL aggregate path, so the driver
+ * answers already-bucketed rows: the main grid (stage × month), the per-stage
+ * subtotal, and the grand total, told apart by what each call groups by.
+ */
+function drillService(): AnalyticsService {
+  return new AnalyticsService({
+    queryCapabilities: () => ({ nativeSql: false, objectqlAggregate: true, inMemory: false }),
+    executeAggregate: async (_object, { groupBy }) => {
+      const grouped = groupedFields(groupBy);
+      if (grouped.includes('close_date')) {
+        return [
+          { stage: 'qualification', close_date: '2026-06', cnt: 3 },
+          { stage: 'won', close_date: '2026-07', cnt: 2 },
+        ];
+      }
+      if (grouped.includes('stage')) {
+        return [
+          { stage: 'qualification', cnt: 3 },
+          { stage: 'won', cnt: 2 },
+        ];
+      }
+      return [{ cnt: 5 }];
+    },
+    getReadScope: (_o, ctx?: ExecutionContext) => (ctx?.tenantId ? { organization_id: ctx.tenantId } : undefined),
+  });
+}
+
+const DRILL_SIDECARS = ['dimensionFields', 'drillRawRows', 'drillRawTotals', 'drillRanges'];
+
+describe('the drill-through sidecars are emitted on the declared AnalyticsResult (#20727)', () => {
+  it('a grouped, date-bucketed answer with totals carries all four, byte for byte', async () => {
+    const answer: AnalyticsResult = await drillService().queryDataset(
+      DATASET,
+      { dimensions: ['stage', 'close_date'], measures: ['cnt'], totals: { groupings: [['stage'], []] } },
+      CTX,
+    );
+
+    // All four are present, in the order the service sets them.
+    expect(Object.keys(answer).filter((k) => DRILL_SIDECARS.includes(k))).toEqual(DRILL_SIDECARS);
+
+    const sidecars = {
+      dimensionFields: answer.dimensionFields,
+      drillRawRows: answer.drillRawRows,
+      drillRawTotals: answer.drillRawTotals,
+      drillRanges: answer.drillRanges,
+    };
+    expect(JSON.stringify(sidecars)).toBe(
+      '{"dimensionFields":{"stage":"stage"},' +
+        '"drillRawRows":[{"stage":"qualification"},{"stage":"won"}],' +
+        '"drillRawTotals":[[{"stage":"qualification"},{"stage":"won"}],[{}]],' +
+        '"drillRanges":[' +
+        '{"close_date":{"field":"close_date","gte":"2026-06-01","lt":"2026-07-01"}},' +
+        '{"close_date":{"field":"close_date","gte":"2026-07-01","lt":"2026-08-01"}}]}',
+    );
+  });
+});
@@ -88,53 +88,6 @@ import { invalidMemberError } from './dataset-refusal.js';
 // it rather than restating it.
 import { ACCEPTED_SQL_DIALECTS, isUnrecognisedSqlDialectAnswer, type AcceptedSqlDialect } from './text-match-sql.js';
 
-/**
- * Analytics result augmented with drill-through metadata (ADR-0021 D2; see
- * queryDataset). Carried alongside `rows` so the host can drill a clicked bucket
- * back to the underlying records without the renderer knowing field mappings.
- *
- * The object those records belong to is not part of this augmentation: it is
- * the contract's own `AnalyticsResult.object`, which every dataset answer
- * carries, drillable or not (#20644).
- */
-type AnalyticsResultWithDrill = AnalyticsResult & {
-  /** Selected drillable dimension NAME → underlying object FIELD name. */
-  dimensionFields?: Record<string, string>;
-  /**
-   * RAW grouped values per row, aligned to `rows` by index — each a map of
-   * drillable dimension NAME → stored value (BEFORE label resolution rewrote
-   * `rows[i][dim]` to the display label). The exact-match drill filter is built
-   * from these, never from the display labels.
-   */
-  drillRawRows?: Array<Record<string, unknown>>;
-  /**
-   * RAW grouped values for the totals/subtotal rows (#3214), the totals-side
-   * companion to `drillRawRows`: `drillRawTotals[i]` aligns to `result.totals[i]`
-   * and `drillRawTotals[i][j]` to `result.totals[i].rows[j]`. Each map holds that
-   * grouping's DRILLABLE dimension NAME → stored value, snapshotted in the SAME
-   * pre-label-resolution pass (the totals loop below overwrites a subtotal row's
-   * dimension value with its display label just like the data rows). Restricted
-   * to the drillable dims present in the grouping, so the grand-total grouping
-   * (`[]`) contributes an empty map per row — which keeps the index alignment
-   * intact and correctly drills the whole (unfiltered) object.
-   */
-  drillRawTotals?: Array<Array<Record<string, unknown>>>;
-  /**
-   * #1752 — half-open date-range drill scope per row, the RANGE companion to
-   * `drillRawRows` (which handles equality dims). A time-bucketed date
-   * dimension (`dateGranularity`) groups a SPAN of records into one bucket
-   * ("2026-Q2"), so its drill needs `[gte, lt)`, not equality — the humanized
-   * bucket can't be exact-matched (which is why date dims are excluded from
-   * `dimensionFields`/`drillRawRows`). Aligned to `rows` by index; each entry
-   * maps a drillable date-dimension NAME → `{ field, gte, lt }` with `gte`
-   * inclusive and `lt` exclusive (bounds as `YYYY-MM-DD`). Present only for
-   * buckets whose boundaries are unambiguous — a `datetime` field under a
-   * non-UTC reference timezone is omitted (host drills an unscoped superset)
-   * until instant-boundary support lands.
-   */
-  drillRanges?: Array<Record<string, { field: string; gte: string; lt: string }>>;
-};
-
 /**
  * [#5717] Does this error carry an ADR-0112 envelope — i.e. did its PRODUCER
  * already classify it?
@@ -1967,10 +1920,10 @@ export class AnalyticsService implements IAnalyticsService {
     // exact-matched against the stored timestamp, so they are not drillable.
     const drillDims = selectedDims.filter((d) => !!d.field && d.type !== 'date');
     if (drillDims.length && result.rows.length) {
-      (result as AnalyticsResultWithDrill).dimensionFields = Object.fromEntries(
+      result.dimensionFields = Object.fromEntries(
         drillDims.map((d) => [d.name, d.field as string]),
       );
-      (result as AnalyticsResultWithDrill).drillRawRows = result.rows.map((row) => {
+      result.drillRawRows = result.rows.map((row) => {
         const raw: Record<string, unknown> = {};
         for (const d of drillDims) raw[d.name] = row[d.name];
         return raw;
@@ -1982,7 +1935,7 @@ export class AnalyticsService implements IAnalyticsService {
       // groups by (the grand-total grouping `[]` keeps empty maps, so a subtotal
       // drill filters by the stored value while the grand total drills unfiltered).
       if (result.totals?.length) {
-        (result as AnalyticsResultWithDrill).drillRawTotals = result.totals.map((total) => {
+        result.drillRawTotals = result.totals.map((total) => {
           const groupingDims = drillDims.filter((d) => total.dimensions.includes(d.name));
           return total.rows.map((row) => {
             const raw: Record<string, unknown> = {};
@@ -2031,7 +1984,7 @@ export class AnalyticsService implements IAnalyticsService {
     if (rangeDims.length && result.rows.length) {
       const bound = (ymd: string, instant: boolean): string =>
         instant ? new Date(zonedDateStartToUtcMs(ymd, rangeTz)).toISOString() : ymd;
-      (result as AnalyticsResultWithDrill).drillRanges = result.rows.map((row) => {
+      result.drillRanges = result.rows.map((row) => {
         const ranges: Record<string, { field: string; gte: string; lt: string }> = {};
         for (const { d, granularity, instant } of rangeDims) {
           // A row in the empty bucket carries `null` here (#3839) and yields no