about summary refs log tree commit diff
path: root/db
diff options
context:
space:
mode:
authorEugen Rochko <eugen@zeonfederated.com>2019-07-29 20:40:21 +0200
committerGitHub <noreply@github.com>2019-07-29 20:40:21 +0200
commite136112ab76a37cbde5c1bcfeb562c8e0a5b5116 (patch)
treedb91ad2e6b5579c6682b4cab0c3488bad506595d /db
parentaefeb65656605f3e32021acbc78c8e21307921fd (diff)
Fix tag normalization and migration not removing duplicate tags (#11441)
Fix #11428
Diffstat (limited to 'db')
-rw-r--r--db/migrate/20190726175042_add_case_insensitive_index_to_tags.rb13
1 files changed, 13 insertions, 0 deletions
diff --git a/db/migrate/20190726175042_add_case_insensitive_index_to_tags.rb b/db/migrate/20190726175042_add_case_insensitive_index_to_tags.rb
index 6fa8c0ec4..057fc86ba 100644
--- a/db/migrate/20190726175042_add_case_insensitive_index_to_tags.rb
+++ b/db/migrate/20190726175042_add_case_insensitive_index_to_tags.rb
@@ -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'