python 44 lines · 7 steps

Caching a Django dashboard aggregate query

A cache-first helper computes a team's 30-day revenue summary with one aggregate query, then stores the result for fifteen minutes.

Explained by highlit

Django Intermediate caching aggregation conditional-queries cache-invalidation

1from django.core.cache import cache
2from django.db.models import Count, Sum, Q
3from django.utils import timezone
4 
5from .models import Order
6 
7 
8def get_revenue_summary(team_id, *, force_refresh=False):
9    cache_key = f"dashboard:revenue_summary:team={team_id}"
10    summary = cache.get(cache_key)
11 
12    if summary is not None and not force_refresh:
13        return summary
14 
15    window_start = timezone.now() - timezone.timedelta(days=30)
16 
17    aggregates = (
18        Order.objects.filter(team_id=team_id, created_at__gte=window_start)
19        .aggregate(
20            gross_revenue=Sum("total", filter=Q(status="paid")),
21            refunded=Sum("total", filter=Q(status="refunded")),
22            paid_orders=Count("pk", filter=Q(status="paid")),
23            open_orders=Count("pk", filter=Q(status="pending")),
24        )
25    )
26 
27    gross = aggregates["gross_revenue"] or 0
28    refunded = aggregates["refunded"] or 0
29 
30    summary = {
31        "gross_revenue": gross,
32        "refunded": refunded,
33        "net_revenue": gross - refunded,
34        "paid_orders": aggregates["paid_orders"],
35        "open_orders": aggregates["open_orders"],
36        "generated_at": timezone.now().isoformat(),
37    }
38 
39    cache.set(cache_key, summary, timeout=60 * 15)
40    return summary
41 
42 
43def invalidate_revenue_summary(team_id):
44    cache.delete(f"dashboard:revenue_summary:team={team_id}")

01 / 01

STEP 01

Walkthrough

Space play ←→ step click any line

Three takeaways

1A stable, parameterized cache key lets you read, write, and invalidate the same value consistently.
2Filtered aggregates compute several conditional totals in a single database round trip.
3Pairing a cache-set with an explicit delete function keeps derived data fresh when the source changes.

Related explainers

java

import java.util.List;
import java.util.IntSummaryStatistics;
 
public class OrderReport {

Aggregating stats with IntSummaryStatistics

streams aggregation records

Intermediate 5 steps

python

import time
from contextlib import contextmanager

A reusable timing context manager in Python

context-manager timing generators

Intermediate 5 steps

typescript

type Fetcher<T> = (key: string) => Promise<T>;
 
export class RequestDeduplicator<T> {
  private inFlight = new Map<string, Promise<T>>();

Deduplicating in-flight requests in TypeScript

deduplication promises caching

Intermediate 7 steps

package cache
 
import (
	"context"

Cache-aside reads with singleflight in Go

caching singleflight concurrency

Advanced 8 steps

python

import os
from datetime import datetime, timezone
 
import jwt

JWT bearer auth as a FastAPI dependency

authentication jwt dependency-injection

Intermediate 7 steps

java

public List<String> collectTagsFromArticles(List<Article> articles) {
    return articles.stream()
        .flatMap(article -> article.getTags().stream())
        .map(String::trim)