Thread: Request for Insights on ID Column Migration Approach

I am just contacting you to talk about a current issue with our database. We have run out of a positive sequence in one of our tables and are now operating with negative sequences. To address this, we plan to migrate from the int4 ID column to an int8 ID column.

The plan involves renaming the int8 column to the id column and setting it as the primary key. However, this process will require downtime, which may be substantial in a production environment. Fortunately, we have noted that other tables do not use the id column as a foreign key, which may help mitigate some concerns.

Our Approach:

1. Copy Data: Copy all the data to the new  id_copy column. 

2. Create a Unique Index: We will then create a unique index on the new ID column before renaming it ,and alter it to non-nullable. This step will necessitate scanning the entire table to verify uniqueness and non-nullability.

3. Add Primary Key: After ensuring the uniqueness, we will add the ID column as the primary key. By doing this, we hope to bypass the additional scanning for uniqueness and nullability, as the column will already be set as not nullable and will have the uniqueness constraint from the unique index.

We want to confirm if this approach will work as expected. If we should be aware of any potential pitfalls or considerations, could you please provide insights or point us toward relevant documentation?

Thank you so much for your help, and I look forward to your guidance.

Best regards,

Aditya Narayan Singh
Loyalty Juggernaut Inc.