Skip to main content

select_related

select_related solves the N+1 query problem by fetching related objects in a single JOIN query.

The Problem​

# N+1 queries!
posts = await Post.objects.filter(active=True)
for post in posts:
    print(post.author.name)  # Each access triggers a new query

The Solution​

from ryx.relations import apply_select_related

# Single query with JOIN
posts = await Post.objects.filter(active=True)
posts = await apply_select_related(posts, fields=["author"])

for post in posts:
    print(post.author.name)  # No extra queries

How It Works​

select_related reconstructs the query with LEFT JOIN:

-- Without select_related
SELECT * FROM "posts" WHERE "active" = ?

-- With select_related(["author"])
SELECT "posts".*, "authors".* FROM "posts"
LEFT JOIN "authors" ON "posts"."author_id" = "authors"."id"
WHERE "posts"."active" = ?

The Rust executor decodes the joined rows and reconstructs full model instances with related objects attached.

Multiple Relations​

posts = await Post.objects.filter(active=True)
posts = await apply_select_related(posts, fields=["author", "category"])
  • ForeignKey relationships — Always works
  • OneToOne relationships — Always works
  • When you know you'll access the related object
  • When the related table is small (JOINs are cheap)
  • ManyToMany relationships — Use prefetch_related instead
  • When you won't access the related object (wasted JOIN)
  • When the related table is huge (JOINs are expensive)

Next Steps​

→ prefetch_related — Eager loading for M2M