Lets say I have a table called users, who looks something like the following:

users table:

|user_id | name | 
   1        Bob   
   2        Alice  
   3        Joe    
   4        Tom   

And another table called books, who got a "users" columns of array of integers relates to users ids who own the books (type: ARRAY[]::integer[])


name     num_pages     users 
mybook1  130           {4,2,1,44}
mybook2  140           {1,2}
mybook3  145           {5,7,8,10}

unfortunately, some of the users has been deleted, and now I have to remove their user id reference from the table.

My question is - Which query should I run in order to remove from the books table the user_ids who are not exists anymore in "users" table? In this case I would like to remove any user id who is not 1/2/3/4 from the users column

for example on "books" table, mybook1 user id 44 does not exist anymore, or mybook3 ids 5,7,8,10 does not exist in users table

(I've created a trigger to handle it from now on when removing users but how can I remove the current "orphan users" who is not reflected in the users table anymore)

note: I have no foreign keys or something like that in the books table

Thanks a lot

  • 1
    This would be a simple DELETE statement if that was properly normalized
    – user1822
    Apr 22, 2019 at 16:04
  • Use a joing table (or associative entity and get rid of arrays. They are a non-relational construct. I often think they're added to relational databases as a kind of "me-too" part of a feature list, rather than a real need - i.e. a need which couldn't have been dealt with relationally.
    – Vérace
    Apr 22, 2019 at 20:16
  • 1
    Because of software application architecture problems , I cant use a joing table or stuff like that.. I just simplified the problem for the question in the current situation
    – Chenko47
    Apr 22, 2019 at 21:02

2 Answers 2


This should perform nicely:

UPDATE books AS b
SET    users = bu.users
   SELECT b1.name, u.users
   FROM   books b1
         SELECT user_id
         FROM   unnest(b1.users) WITH ORDINALITY u(user_id, ord)
         JOIN   users u USING (user_id)
         ORDER  BY u.ord
      ) u(users)
   ) bu
WHERE   b.name = bu.name
AND     b.users <> bu.users;

Preserves original order of array elements. If you don't need to, drop ORDINALITY & friends from the query.

Can result in empty arrays.

Assuming ...

  • books.name is the PK. Else use some other UNIQUE column from books instead.
  • users.user_id is at least UNIQUE.
  • books.users does not contain NULL values. Else those are eliminated as well.
CREATE OR REPLACE FUNCTION intersection(anyarray, anyarray) RETURNS anyarray as $$
                    SELECT ARRAY(
                    SELECT $1[i]
                    FROM generate_series( array_lower($1, 1), array_upper($1, 1) ) i
                    WHERE ARRAY[$1[i]] && $2
                    $$ language sql;

                    user_ids INTEGER[];
                    row books%ROWTYPE;
                    user_ids := (SELECT array_agg(id) FROM users) ;

                    FOR row IN
                        SELECT * FROM books
                        UPDATE books
                        SET users=intersection(row.users,user_ids) 
                        WHERE id = row.id;
                    END LOOP;
            $BODY$ language plpgsql

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.