VladPolskiy/gitlab-foss
1
0
Fork 0
mirror of https://gitlab.com/gitlab-org/gitlab-foss.git synced 2025-07-25 16:03:48 +00:00
Code Issues Packages Projects Releases 1 Wiki Activity

Add latest changes from gitlab-org/gitlab@master

Browse Source
This commit is contained in:
GitLab Bot
2025-07-19 03:10:37 +00:00
parent 23158f6b5f
commit e1d81b823f
2 changed files with 22 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
lib/gitlab/database/repair_index.rb
View File

@ -15,6 +15,7 @@ module Gitlab
FIND_DUPLICATE_SETS_SQL = <<~SQL FIND_DUPLICATE_SETS_SQL = <<~SQL
SELECT ARRAY_AGG(id ORDER BY id ASC) as ids SELECT ARRAY_AGG(id ORDER BY id ASC) as ids
FROM %{table_name} FROM %{table_name}
WHERE %{not_null_conditions}
GROUP BY %{column_list} GROUP BY %{column_list}
HAVING COUNT(*) > 1 HAVING COUNT(*) > 1
SQL SQL
@ -323,10 +324,15 @@ module Gitlab
def find_duplicate_sets(table_name, columns) def find_duplicate_sets(table_name, columns)
logger.info("Checking for duplicates in '#{table_name}' for columns: #{columns.join(',')}...") logger.info("Checking for duplicates in '#{table_name}' for columns: #{columns.join(',')}...")
not_null_conditions = columns.map do |col|
"#{connection.quote_column_name(col)} IS NOT NULL"
end.join(' AND ')
sql = format( sql = format(
FIND_DUPLICATE_SETS_SQL, FIND_DUPLICATE_SETS_SQL,
table_name: connection.quote_table_name(table_name), table_name: connection.quote_table_name(table_name),
column_list: columns.map { |col| connection.quote_column_name(col) }.join(', ') column_list: columns.map { |col| connection.quote_column_name(col) }.join(', '),
not_null_conditions: not_null_conditions
) )
execute_local(sql, read_only: true) do execute_local(sql, read_only: true) do

21
spec/lib/gitlab/database/repair_index_spec.rb
View File

@ -74,8 +74,8 @@ RSpec.describe Gitlab::Database::RepairIndex, feature_category: :database do
connection.execute(<<~SQL) connection.execute(<<~SQL)
CREATE TABLE #{test_table} ( CREATE TABLE #{test_table} (
id serial PRIMARY KEY, id serial PRIMARY KEY,
name varchar(255) NOT NULL, name varchar(255) NULL,
email varchar(255) NOT NULL email varchar(255) NULL
); );
SQL SQL
@ -184,7 +184,9 @@ RSpec.describe Gitlab::Database::RepairIndex, feature_category: :database do
INSERT INTO #{test_table} (name, email) VALUES INSERT INTO #{test_table} (name, email) VALUES
('test_user', 'test@example.com'), -- ID 1 ('test_user', 'test@example.com'), -- ID 1
('test_user', 'test@example.com'), -- ID 2 (duplicate) ('test_user', 'test@example.com'), -- ID 2 (duplicate)
('other_user', 'other@example.com'); -- ID 3 ('test_user', NULL), -- ID 3, email NULL, should be preserved
(NULL, 'other@example.com'), -- ID 4, name NULL, should be preserved
('other_user', 'other@example.com'); -- ID 5
SQL SQL
# Create standard references (no entity column) # Create standard references (no entity column)
@ -217,7 +219,7 @@ RSpec.describe Gitlab::Database::RepairIndex, feature_category: :database do
it 'handles all reference types correctly' do it 'handles all reference types correctly' do
# before: 3 users, various references # before: 3 users, various references
user_count_before = connection.select_value("SELECT COUNT(*) FROM #{test_table}") user_count_before = connection.select_value("SELECT COUNT(*) FROM #{test_table}")
expect(user_count_before).to eq(3) expect(user_count_before).to eq(5)
# unique index doesn't exist yet # unique index doesn't exist yet
index_exists_before = connection.select_value(<<~SQL).present? index_exists_before = connection.select_value(<<~SQL).present?
@ -230,9 +232,16 @@ RSpec.describe Gitlab::Database::RepairIndex, feature_category: :database do
repairer.run repairer.run
# after: 2 users (duplicate removed) # after: 4 users (only true duplicate ID 2 removed)
# ID 3 with NULL value preserved
user_count_after = connection.select_value("SELECT COUNT(*) FROM #{test_table}") user_count_after = connection.select_value("SELECT COUNT(*) FROM #{test_table}")
expect(user_count_after).to eq(2) expect(user_count_after).to eq(4)
# Verify NULL values are preserved
null_records = connection.select_value(
"SELECT COUNT(*) FROM #{test_table} WHERE email IS NULL or name is NULL"
)
expect(null_records).to eq(2)
# standard reference updated to good ID # standard reference updated to good ID
standard_ref = connection.select_value( standard_ref = connection.select_value(