CODEWITHSUNDEEP
rubyruby-on-rails-3benchmarking
Jonathan
Jonathan

Reputation: 11494

Two (seemingly) identical queries, one is faster, why?

Two seemingly identical queries (as far as a newbie like me can tell, but the first is faster overall in the partial template rendering time (nothing else changed but the ids statement). Also, when testing through rails console, the latter will visibly run a query, the former will not. I do not understand why - and why the first statement is a few ms faster than the second - though I can guess it is due to the shorter method chaining to get the same result.

UPDATE: My bad. They are not running the same query, but it still is interesting how a select on all columns is faster than a select on one column. Maybe it is a negligible difference compared to the method chaining though.

ids = current_user.activities.map(&:person_id).reverse
SELECT "activities".* FROM "activities" WHERE "activities"."user_id" = 1
SELECT "people".* FROM "people" WHERE "people"."id" IN (1, 4, 12, 15, 3, 14, 17, 10, 5, 6) Rendered activities/_activities.html.haml (7.4ms)


ids = current_user.activities.order('id DESC').select{person_id}.map(&:person_id)
SELECT "activities"."person_id" FROM "activities" WHERE "activities"."user_id" = 1 ORDER BY id DESC    
SELECT "people".* FROM "people" WHERE "people"."id" IN (1, 4, 12, 15, 3, 14, 17, 10, 5, 6) Rendered activities/_activities.html.haml (10.3ms)

The purpose of the statement is to retrieve the foreign key reference to people in the order in which they appeared in the activities table, (on its PK). Note: I use Squeel for SQL.

Upvotes: 0

Views: 101

Answers (1)

A Person
A Person

Reputation: 1350

In the first query, you've chained .map and .reverse, while in the second query, you've used .order('id DESC') .select(person_id) which were unnecessary, if you added .reverse

Upvotes: 1

Related Questions