Fix tag normalization and migration not removing duplicate tags (#11441)

Fix #11428
This commit is contained in:
Eugen Rochko 2019-07-29 20:40:21 +02:00 committed by GitHub
parent aefeb65656
commit e136112ab7
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 51 additions and 4 deletions

View file

@ -2,6 +2,19 @@ class AddCaseInsensitiveIndexToTags < ActiveRecord::Migration[5.2]
disable_ddl_transaction!
def up
Tag.connection.select_all('SELECT string_agg(id::text, \',\') AS ids FROM tags GROUP BY lower(name) HAVING count(*) > 1').to_hash.each do |row|
canonical_tag_id = row['ids'].split(',').first
redundant_tag_ids = row['ids'].split(',')[1..-1]
safety_assured do
execute "UPDATE accounts_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM accounts_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.account_id = t0.account_id)"
execute "UPDATE statuses_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM statuses_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.status_id = t0.status_id)"
execute "UPDATE featured_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM featured_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.account_id = t0.account_id)"
end
Tag.where(id: redundant_tag_ids).in_batches.delete_all
end
safety_assured { execute 'CREATE UNIQUE INDEX CONCURRENTLY index_tags_on_name_lower ON tags (lower(name))' }
remove_index :tags, name: 'index_tags_on_name'
remove_index :tags, name: 'hashtag_search_index'