RE: auto generated ID as primary key

AFAIK, having an arbitrary ID identity column as primary key is considered best practice.



-----Original Message-----
From: Jay Vaughn [mailto:jeffersonvaughn@xxxxxxxxx]
Sent: Thursday, February 01, 2018 7:39 AM
To: midrange-l@xxxxxxxxxxxx
Subject: auto generated ID as primary key

though i know this is bad practice (to have keys defined in tables), my current employer currently has tables with...

L_ID for column ID
BIGINT not null generated always as
identity (start with 1 increment by 1),

Now when this file is "re-homed" it knocks the auto generate row id out of whack and creates a duplicate record error when a pgm tries to write to it. And of course we have to manually reset the row id.

My guess is because of the (start with 1 increment by 1) - meaning, once it is re-homed it somehow thinks it needs to start at 1 again.

Is that true? Ideally i'd like to remove the ID from being a primary key on the table, but would at least removing (start with 1 increment by 1) resolve the re-homing and duplicate record issue, or no? How exactly does this work?