Optional dashboard filters
Reuse a query while adding filters supplied by an application.
Keep required tenant and date filters in the base query, then add an event filter when one is supplied. Start with the shared setup.
Query
export function filteredCounts(
db: AnalyticsDB,
tenantId: number,
filters: { from: string; eventType?: string },
) {
let query = db
.selectFrom("events")
.prewhere("tenant_id", "=", tenantId)
.where("created_at", ">=", filters.from);
if (filters.eventType !== undefined) {
query = query.where("event_type", "=", filters.eventType);
}
return query
.selectExpr((eb) => ["event_type", eb.fn.count().as("events")])
.groupBy("event_type")
.orderBy("event_type");
}Builders are immutable: where() returns a new query. Assigning it back to
query is what makes the optional filter take effect. This branch changes only
a predicate, so both paths retain the same result shape.
const query = filteredCounts(db, 1, {
from: "2026-08-04 00:00:00",
eventType: "purchase",
});
const rows = await query.execute();
// { event_type: string; events: string }[]SQL
SELECT event_type, count() AS events
FROM events
PREWHERE tenant_id = {p0:Int64}
WHERE created_at >= {p1:String} AND event_type = {p2:String}
GROUP BY event_type
ORDER BY event_type ASCParameters: p0 = 1, p1 = "2026-08-04 00:00:00", p2 = "purchase".
The values travel separately from the SQL text.
Result
[{ "event_type": "purchase", "events": "1" }]The quickstart's src/server.ts validates dates and allowed event types before
calling its query function. Apply your own request validation at that boundary
and resolve the tenant from the authenticated caller. Quarry's schema types
check query construction; they do not validate an incoming HTTP request.
To cancel a query when the request ends, pass its signal:
await query.execute({ abortSignal: requestSignal });Cancellation requires Quarry 0.10.0 or newer; see release guidance. For reusable expressions and unions, continue with query composition.