As you can see in the current code below, I am finding the duplicate based on the attribute recordable_id. What I need to do is find the duplicate based on four matching attributes: user_id, recordable_type, hero_type, recordable_id. How must I modify the code?
heroes = User.heroes
for hero in heroes
hero_statuses = hero.hero_statuses
seen = []
hero_statuses.sort! {|a,b| a.created_at <=> b.created_at } # sort by created_at
hero_statuses.each do |hero_status|
if seen.map(&:recordable_id).include? hero_status.recordable_id # check if the id has be开发者_运维问答en seen already
hero_status.revoke
else
seen << hero_status # if not, add it to the seen array
end
end
end
Try this:
HeroStatus.all(:group => "user_id, recordable_type, hero_type, recordable_id",
:having => "count(*) > 1").each do |status|
status.revoke
end
Edit 2 To revoke the all the latest duplicate entries do the following:
HeroStatus.all(:joins => "(
SELECT user_id, recordable_type, hero_type,
recordable_id, MIN(created_at) AS created_at
FROM hero_statuses
GROUP BY user_id, recordable_type, hero_type, recordable_id
HAVING COUNT(*) > 1
) AS A ON A.user_id = hero_statuses.user_id AND
A.recordable_type = hero_statuses.recordable_type AND
A.hero_type = hero_statuses.hero_type AND
A.recordable_id = hero_statuses.recordable_id AND
A.created_at < hero_statuses.created_
").each do |status|
status.revoke
end
Using straight Ruby (not the SQL server):
heroes = User.heroes
for hero in heroes
hero_statuses = hero.hero_statuses
seen = {}
hero_statuses.sort_by!(&:created_at)
hero_statuses.each do |status|
key = [status.user_id, status.recordable_type, status.hero_type, status.recordable_id]
if seen.has_key?(key)
status.revoke
else
seen[key] = status # if not, add it to the seen array
end
end
remaining = seen.values
end
For lookups, always use Hash
(or Set
, but here I thought it would be nice to keep the statuses that have been kept)
Note: I used sort_by!
, but that's new to 1.9.2, so use sort_by
(or require "backports"
)
精彩评论