How to update each row in table with random row from another table?

Question

I am doing some testing and need to populate Employee.DepartmentId with Id column from Department table.

When I do like this, All rows in Employee table have same DepartmentId value. But I want different DepartmentId values, it is ok if there are some duplicates.

UPDATE Employee SET DepartmentId = (SELECT Id FROM Department ORDER BY RANDOM() LIMIT 1);

How can I do this?

Adding more details :

How can I write SQL query so that for each row in Employee , engine goes and fetches a random row from Department table, rather than fetch one random row from Department table and use that for every row in Employee .

Answer 1

This is a problem of over-active optimizers which decide to run the subquery only once.

One solution is to add a correlation clause to the outer query:

UPDATE Employee
    SET DepartmentId = (SELECT d.Id
                        FROM Department d
                        WHERE employee.id is not null
                        ORDER BY RANDOM() LIMIT 1
                       );

This gets around the optimizer.

Here is a db<>fiddle.