Skip to content

Automatic translation — the Portuguese version is the reference. View in Portuguese

Updated on 2026-10-04

Filters and interactions

Dashboard filters, cross-filtering between visuals and what happens when you click a data point.

On this page (8)

Key finding: every filter/slicer is client-side on top of the snapshot — it never becomes SQL. That has three good consequences: zero cost per interaction, no injection surface through slicers, and the limit becomes what's in the snapshot. Examples use the synthetic workspace Comércio Aurora Ltda.


Slicers (8 types) and scopes#

  1. What it is — 8 slicer types: select, daterange, buttons, range, hier (cascading), number ("what if"), field (field parameter) and measure.
  2. When to use it — slice the dashboard by category, period, range; or swap field/measure (field/measure) and simulate a number (number/what-if).
  3. How it works — each slicer has a scope: dashboard, page or per visual (target); search in the filter (≥20 options), multi-select (OR/AND), select all/clear. Sync across pages (syncPages): a control on one page applies to the whole dashboard.
  4. Prerequisites — a snapshot with the columns used by the slicer.
  5. How to validate — add a select on produtos.categoria scoped to the dashboard, select "Bebidas" and confirm every visual slices; change the scope to a single visual and confirm the warning when it "doesn't reach anything".
  6. Limits and caveats — personal per-user filters (personal slicers) are Roadmap; what exists per user is RLS (see Security). Slicers operate on the snapshot, not on the live source.
  7. Sample data (synthetic) — categories {Bebidas, Limpeza, Higiene}; selecting "Bebidas" leaves only revenue=250.
  8. Evidence (code/test) — DASH_FILTER_TYPES in dashboard-widgets.ts; renders in dashboard-filters.tsx; slicer tests.

Relative dates#

  1. What it is — period presets: today, 7d, 30d, month, year, and the generic -Nd/-Nw/-Nm/-Ny (Power BI style), in the dashboard's time zone.
  2. When to use it — a "last 30 days" that moves forward on its own every day.
  3. How it works — presetRange calculates the window in the dashboard's configured time zone.
  4. Prerequisites — a date column in the snapshot.
  5. How to validate — set "last 30 days" and confirm the range shifts when you change the reference date.
  6. Limits and caveats — it depends on the dashboard's time zone (America/Sao_Paulo by default).
  7. Sample data (synthetic) — today=2026-07-31 → window 2026-07-02..07-31.
  8. Evidence (code/test) — presetRange (dashboard-widgets.ts) + test.

Cross-filter / cross-highlight and the interactions editor#

  1. What it is — clicking a visual highlights the selection and slices the others; plus a source→target interactions editor to define, per pair of visuals, whether the action is Filter / Highlight / Ignore.
  2. When to use it — the gesture that "feels like Power BI": click a bar and see the rest of the dashboard react.
  3. How it works — a click applies cross-filtering (Ctrl accumulates; clicking again clears); the pairs editor sets the mode per target, inheriting the global mode ("default") with an in-memory preview before saving; sanitization with an allowlist (valid modes, existing ids, no self-reference, a ceiling on pairs). The same semantics apply in the authenticated viewer, the public view and embeds.
  4. Prerequisites — two or more visuals on the page.
  5. How to validate — in the interactions editor, set "Category bars → Table = Filter"; click a bar and confirm the table slices; switch to "Ignore" and confirm the table doesn't react.
  6. Limits and caveats — persistence of the selection across sessions is partial (see below).
  7. Sample data (synthetic) — clicking "Bebidas" on the bars filters the table to the Bebidas items.
  8. Evidence (code/test) — interactions-editor.tsx, interactions-sanitize.ts + test; applyCrossFilter (dashboard-widgets.ts); dashboard-view.tsx.

Drill-down / drill-up#

  1. What it is — going down/up levels through a column hierarchy and through dates (year → month → day).
  2. When to use it — exploring from "Region" down to "City", or from year to day.
  3. How it works — config.drill over the hierarchy; context chips + a drill-up button; dates use dateBucket.
  4. Prerequisites — a hierarchy defined in the model (see Semantic model).
  5. How to validate — on a visual with the hierarchy Região > UF > Cidade, click to go down from Sudeste to São Paulo and use drill-up to go back.
  6. Limits and caveats — drill operates on the visual's snapshot.
  7. Sample data (synthetic) — Sudeste → SP → São Paulo.
  8. Evidence (code/test) — config.drill (dashboard-widgets.ts), dashboard-filters.tsx.

Drill-through between pages#

  1. What it is — jumping to another page carrying the context (by column), automatically on click or via a button, with a "back" chip.
  2. When to use it — click a product and go to that product's detail page.
  3. How it works — DashPageTab.drillthrough defines the key column; the target opens filtered by the clicked value.
  4. Prerequisites — a target page configured as a drill-through target.
  5. How to validate — set "Product detail" as the target by produtos.id; click a product in the overview and confirm the detail page is already filtered.
  6. Limits and caveats — the context passed is per configured column.
  7. Sample data (synthetic) — clicking product id=10 opens "Detail" with produto_id=10.
  8. Evidence (code/test) — DashPageTab.drillthrough, dashboard-filters.tsx.

Tooltips and tooltip pages#

  1. What it is — tooltips with extra fields, rich tooltips (content beyond the number) and tooltip pages (a whole dashboard page used as a tooltip, Power BI style).
  2. When to use it — showing details on hover without cluttering the visual.
  3. How it works — config.tooltipFields/tooltipPage; the tooltip page is hidden from the tab bar and appears when you hover over the point.
  4. Prerequisites — extra fields and/or a page marked as a tooltip.
  5. How to validate — add produtos.marca as a tooltip field and confirm it appears on hover; create a tooltip page and confirm the preview.
  6. Limits and caveats — a rich tooltip is structured content; the data engine is the same as the source visual's.
  7. Sample data (synthetic) — a "Revenue by category" tooltip showing brand and number of orders.
  8. Evidence (code/test) — dashboard-tooltip.ts (+15 tests), tooltip-rich.ts + test.

Bookmarks#

  1. What it is — up to 10 bookmarks that save the dashboard's complete state, applicable via a button.
  2. When to use it — saving an "executive view" and a "detailed view" and switching with one click.
  3. How it works — the bookmark saves selections/filters/state; a button can apply it.
  4. Prerequisites — a configured state to save.
  5. How to validate — set filters, save as a bookmark "July", change everything and apply the bookmark to restore.
  6. Limits and caveats — bookmarks belong to the dashboard, not to each user.
  7. Sample data (synthetic) — bookmark "Bebidas · July".
  8. Evidence (code/test) — saveBookmarksAction, dashboard-filters.tsx.

Persisting and sharing selections via URL#

  1. What it is — the selection travels in the URL (?f=, base64url) so you can share a filtered view.
  2. When to use it — sending an already-filtered link to a colleague.
  3. How it works — the URL stores selection + date; parameters (field/measure/what-if) also travel through the state codec.
  4. Prerequisites — filters/parameters applied.
  5. How to validate — filter "Bebidas", copy the URL, open it in another tab and confirm the filter is restored.
  6. Limits and caveats — range/hierarchy don't survive through the URL alone (they depend on a bookmark); there's no per-user autosave.
  7. Sample data (synthetic) — ?f=… with category=Bebidas, period=July.
  8. Evidence (code/test) — dashboard-filters.tsx (?f=), url-state.ts.

Related links