I have 11 tables in a single schema in PostgreSQL. Each table has a column named 'id' that should be a unique primary key. Unfortunately, some of the tables have ids that are in other tables.
How can I create a new sequence and update all the tables' id values with fully unique values? Do I have to do it one sequence and one table at a time?
I'm new at SQL, so I don't know where to start beyond creating a sequence and updating a single table.
Your question is a bit vague but I assume your ids are of type int/bigint. And you want these 'id's to be globally unique not just inside one table.
There might be many options. This one is one of the most simple but maybe not the most elegant way. First create a sequence to provied ids, like:
CREATE SEQUENCE globally_unique_id;
Then just go through all your 11 tables with update, like:
UPDATE table1 SET id = nextval('globally_unique_id');
Using this one sequence will not give you the same id for any rows in all tables.
If you have foreign key constraints in other tables to the 'id' column that are not declared to cascade on update you are in bigger trouble. But that is an another story.
Of course you do not need to do it table by table. You could for example query Postgres system tables for all tables - like asked here - (or make an array of table names by hand) and loop tables with id in some specially constructed query or procedure.
But in my opinion: 'only' 11 tables might not be worth of coding it. On the other hand it might be useful later if you need to do the same again.
