ActiveRecord includes and preload

me@iany.me (Ian Yang) — Fri, 23 Aug 2013 00:00:00 +0000

ActiveRecord has two query methods to eager load associations, includes and preload. Although the documentation of preload says

Allows preloading of args, in the same way that includes does.

Indeed the two methods have some differences.

In simple words, prefer includes to eager load associations. Use preload only when

you want to customize select columns, or
you meet error “Can not eagerly load the polymorphic association”.

Both of the two methods eager load associations to avoid N+1 queries. However, includes is more aggressive. If it detects that the associations tables are already joined, through explicitly joins or implicit references in where, it overwrites select fields from the joined query and construct result set from it, rather than loading associations in a separate query. See the generated SQL statement below and the fields aliases like t0_r0.

irb(main):049:0> User.includes(:articles).joins(:articles)
SQL (0.7ms) SELECT "users"."id" AS t0_r0,
"users"."login" AS t0_r1,
"users"."name" AS t0_r2,
"users"."created_at" AS t0_r3,
"users"."updated_at" AS t0_r4,
"articles"."id" AS t1_r0,
"articles"."title" AS t1_r1,
"articles"."content" AS t1_r2,
"articles"."user_id" AS t1_r3,
"articles"."created_at" AS t1_r4,
"articles"."updated_at" AS t1_r5
FROM "users"
INNER JOIN "articles"
ON "articles"."user_id" = "users"."id"

However, if the table is not joined yet, includes will fallback to preload, which loads association in a separate query. See the SQL in log below that articles are eager loaded in another SELECT query.

irb(main):050:0> User.includes(:articles)
User Load (0.1ms) SELECT "users".* FROM "users"
Article Load (0.2ms) SELECT "articles".* FROM "articles"
WHERE "articles"."user_id" IN (1, 2)

Because includes may (or may not) overwrite select, if you have your own select clause, use preload instead. See example below:

users = User.joins(:articles).select(
'users.*, articles.created_at as last_posted_at'
)
users.includes(:articles).first.last_posted_at
# NoMethodError: undefined method `last_posted_at'
users.preload(:articles).first.last_posted_at
# => "2013-08-23 11:20:36.536968"

Older version of Rails has trouble to includes polymorphic associations. It seems the newer version of Rails is smart enough to fallback to preload. However if you still see the error “Can not eagerly load the polymorphic association”, try switching to preload.

References

ActiveRecord uniq, count and distinct

me@iany.me (Ian Yang) — Tue, 16 Jul 2013 00:00:00 +0000

ActiveRecord has two methods to remove duplicates. Method uniq and option distinct: true in method count. I thought uniq.count and count(distinct: true) were identical. Indeed, uniq.count still counts duplicates, and count(distinct: true) must be used here.

In simple words, use uniq to get unique result set, use count(distinct: true) to count unique result.

For example, user has many activities, and I want to get all users having a specific type of activities:

users = User.joins(:activities).where(
activities: { activity_type: 'purchase'}
)

Because a user may have multiple activities with the same type, the result above may contain duplicate users. Method uniq can be used here to remove the duplicates:

users = users.uniq

But users.uniq.count generates SQL like below:

SELECT DISTINCT COUNT(*) ...

This SQL counts all records with duplicates, and apply DISTINCT on the count, which has only one row. So DISTINCT has no effect here.

On the other hand, users.count(distinct: true) generates SQL below, which removes duplicates first, then count the result.

SELECT COUNT(DISTINCT users.id) ...

~iany/ Active Record

ActiveRecord includes and preload

References

ActiveRecord uniq, count and distinct