Which transformers can be used?

have a look at the SQLExecuter. You should be able to add this after your initial write to then extract erroneous data in the DB along with the attributes to populate your errors table

Thank you! I have connected SQL Executor after the Writer and it indeed extracts “ID” field, but for each record it gives all set of ID’s. Struggle to understand how to set SQL query properly, in order ID was extracted specifically for the objects 1 to 1 that were written to the DB. It is my current query:
select "Id" from "schema"."table"
As a result, if I have 12 objects written in the table and DB generated 12 IDs for them, when I run SQLExecutor, I am getting list of 144 records that are repeating.

SQLExecutor will perform your query once for every feature that enters the SQLExecutor.

So when 12 features enter the SQLExecutor, and your query returns 12 records, you will get 12 times 12 = 144 records (12 records, all repeated 12 times).

Solution: change the query so it returns only the corresponding record for each feature that enters the SQLExecutor.

Thank you for for your answer. Yes, I understand this what happens. I don’t understand how to modify the query so SQLexecutor would provide 1 ID per feature. 

Perhaps something like 

Select ID from TABLE where ID = ID_from_Feature

Or maybe something like
Select distinct ID from TABLE 

The problem is that I don’t have ID_from_Feature. The main reason why I need to get assigned by the database ID is to have unique identifier for written objects.

After you have written the features to the database table, surely there must be some way to identify which record corresponds to which feature?

Yes, you are right. We came up to this with dev team that there is a need to add additional field in the DB, otherwise it is hardly possible. Thank you