Sequences to generate primary keys

Question

Hi,
A question on using sequences to generate primary keys. We have a table and in all environments, test, uat, prod, a sequence exists to generate the primary key for the table. This table is a reference table and is not updateable by the users.
When developing scripts to insert records into the table, the developers are currently not using the sequence but typing in unique numbers so data scripts consist of something like this:
INSERT INTO EMPLOYEE_GROUP
(employee_group_id, code, description,...)
VALUES
(45,'sdf','sdfsdfsf');
INSERT INTO EMPLOYEE_GROUP
(employee_group_id, code, description,...)
VALUES
(46,'sdf','sdfsdfsf',...);
a) Is this normal practice?
b) Sometimes they are using the sequence to generate the primary key in development but then hardcoding the number in the script so in development the sequence might be at 160, but in the other environments the sequences will still be at one. Is this also normal practice?
My usual method in developing scripts is to use <seqeunce>.nextval but it's somewhat incompatitble in the current environment.

Answer 1

You are using a surrogate, or synthetic primary key. For a primary key generated in this way, the value is inconsequential. All that matters is that it is unique and the tables with foreign keys into it, if any, also have that value set properly. But it doesn't matter if the value is 1, 100, 12344, 2222222222, etc. All you care about is the uniqueness.
If you create the data properly then child records and their parents should be correct in all systems you mention, it's just that the actual value of the key will differ and that's fine. You care about the relationship of the records, not the value of the synthetic keys.