Composite foreign keys with an incorrect order of columns (ver 1)

Query goal:

Find composite foreign keys where the order of columns does not correspond to the order of columns in the referenced candidate key. Find composite foreign keys in case of which the foreign key and candidate key consist of columns with the same name but the order of columns in the keys is different. For instance, the query returns information about a foreign key (personal_code, country_code) that refers to the candidate key (country_code, personal_code). In SQL keys are ordered sets of columns. Thus, in case of composite foreign key declarations one has to pay attention that the order of columns in the FOREIGN KEY clause matches the order of columns in the REFERENCES clause.

Notes about the query:

There could be a false negative result if the column names in a foreign key and its referenced candidate key are different. For instance, FOREIGN KEY (person_id, car_id) REFERENCES (car_code, person_id) will not be detected.

Query type:

Problem detection (Each row in the result could represent a flaw in the design)

Query reliability:

Medium (Medium number of false-positive results)

Query license:

MIT License

Fixing suggestion:

Drop the foreign key constraint and recreate with the correct column order.

Data source:

system catalog only

SQL query:

Click on query to copy it

with fk as (select 
o.conname,
(select nspname from pg_namespace where oid=f.relnamespace) as foreign_schema,
f.relname as foreign_table,
f.oid as foreign_table_oid,
o.confkey AS foreign_col,
(select nspname from pg_namespace where oid=c.relnamespace) as target_schema,
c.relname as target_table, 
c.oid as target_table_oid,
o.conkey AS target_col
from pg_constraint o inner join pg_class c on c.oid = o.conrelid
inner join pg_class f on f.oid = o.confrelid 
where o.contype = 'f' and cardinality(o.conkey)>1),
fk_unnest as (select conname, foreign_schema, foreign_table,  foreign_table_oid, foreign_col, foreign_col_num, target_schema, target_table, target_table_oid, target_col, target_col_num, ordin
from fk, unnest(fk.foreign_col, fk. target_col) with ordinality as f(foreign_col_num, target_col_num, ordin)),
fk_with_names as (select conname, foreign_schema, foreign_table, array_agg(a_foreign.attname order by ordin) as foreign_col, target_schema, target_table, array_agg(a_target.attname order by ordin) as target_col
from fk_unnest fk inner join pg_attribute a_foreign on fk.foreign_col_num = a_foreign.attnum and fk.foreign_table_oid = a_foreign.attrelid and a_foreign.attisdropped = false
inner join pg_attribute a_target on fk.target_col_num = a_target.attnum and fk.target_table_oid = a_target.attrelid and a_target.attisdropped = false
group by conname, foreign_schema, foreign_table, target_schema, target_table)
select conname, foreign_schema, foreign_table, foreign_col, target_schema, target_table, target_col
from fk_with_names
where foreign_col<>target_col and foreign_col <@ target_col and foreign_col @> target_col
order by target_schema, target_table, conname;

SQL statements for generating SQL statements that help us to fix the problem

SQL query	Description
with fk as (select o.conname, (select nspname from pg_namespace where oid=f.relnamespace) as foreign_schema, f.relname as foreign_table, f.oid as foreign_table_oid, o.confkey AS foreign_col, (select nspname from pg_namespace where oid=c.relnamespace) as target_schema, c.relname as target_table, c.oid as target_table_oid, o.conkey AS target_col from pg_constraint o inner join pg_class c on c.oid = o.conrelid inner join pg_class f on f.oid = o.confrelid where o.contype = 'f' and cardinality(o.conkey)>1), fk_unnest as (select conname, foreign_schema, foreign_table, foreign_table_oid, foreign_col, foreign_col_num, target_schema, target_table, target_table_oid, target_col, target_col_num, ordin from fk, unnest(fk.foreign_col, fk. target_col) with ordinality as f(foreign_col_num, target_col_num, ordin)), fk_with_names as (select conname, foreign_schema, foreign_table, array_agg(a_foreign.attname order by ordin) as foreign_col, target_schema, target_table, array_agg(a_target.attname order by ordin) as target_col from fk_unnest fk inner join pg_attribute a_foreign on fk.foreign_col_num = a_foreign.attnum and fk.foreign_table_oid = a_foreign.attrelid and a_foreign.attisdropped = false inner join pg_attribute a_target on fk.target_col_num = a_target.attnum and fk.target_table_oid = a_target.attrelid and a_target.attisdropped = false group by conname, foreign_schema, foreign_table, target_schema, target_table) SELECT format('ALTER TABLE %1$I.%2$I DROP CONSTRAIN %3$I;', target_schema, target_table, conname) AS statements from fk_with_names where foreign_col<>target_col and foreign_col <@ target_col and foreign_col @> target_col order by target_schema, target_table, conname;	Drop the foreign key constraint.

SQL query

Description

with fk as (select 
o.conname,
(select nspname from pg_namespace where oid=f.relnamespace) as foreign_schema,
f.relname as foreign_table,
f.oid as foreign_table_oid,
o.confkey AS foreign_col,
(select nspname from pg_namespace where oid=c.relnamespace) as target_schema,
c.relname as target_table, 
c.oid as target_table_oid,
o.conkey AS target_col
from pg_constraint o inner join pg_class c on c.oid = o.conrelid
inner join pg_class f on f.oid = o.confrelid 
where o.contype = 'f' and cardinality(o.conkey)>1),
fk_unnest as (select conname, foreign_schema, foreign_table,  foreign_table_oid, foreign_col, foreign_col_num, target_schema, target_table, target_table_oid, target_col, target_col_num, ordin
from fk, unnest(fk.foreign_col, fk. target_col) with ordinality as f(foreign_col_num, target_col_num, ordin)),
fk_with_names as (select conname, foreign_schema, foreign_table, array_agg(a_foreign.attname order by ordin) as foreign_col, target_schema, target_table, array_agg(a_target.attname order by ordin) as target_col
from fk_unnest fk inner join pg_attribute a_foreign on fk.foreign_col_num = a_foreign.attnum and fk.foreign_table_oid = a_foreign.attrelid and a_foreign.attisdropped = false
inner join pg_attribute a_target on fk.target_col_num = a_target.attnum and fk.target_table_oid = a_target.attrelid and a_target.attisdropped = false
group by conname, foreign_schema, foreign_table, target_schema, target_table)
SELECT format('ALTER TABLE %1$I.%2$I DROP CONSTRAIN %3$I;', target_schema, target_table, conname) AS statements
from fk_with_names
where foreign_col<>target_col and foreign_col <@ target_col and foreign_col @> target_col
order by target_schema, target_table, conname;

Drop the foreign key constraint.

Collections where the query belongs to

Collection name	Collection description
Find problems automatically	Queries, that results point to problems in the database. Each query in the collection produces an initial assessment. However, a human reviewer has the final say as to whether there is a problem or not .

Collection name

Collection description

Find problems automatically

Queries, that results point to problems in the database. Each query in the collection produces an initial assessment. However, a human reviewer has the final say as to whether there is a problem or not .

Categories where the query belongs to

Category name	Category description
Relationships between tables	Queries of this category provide information about how database tables are connected to each other and whether such connections have been explicitly defined and whether it has been done correctly.
Result quality depends on names	Queries of this category use names (for instance, column names) to try to guess the meaning of a database object. Thus, the goodness of names determines the number of false positive and false negative results.

Category name

Category description

Relationships between tables

Queries of this category provide information about how database tables are connected to each other and whether such connections have been explicitly defined and whether it has been done correctly.

Result quality depends on names

Queries of this category use names (for instance, column names) to try to guess the meaning of a database object. Thus, the goodness of names determines the number of false positive and false negative results.