The last time I tried to do that, I struggled a lot with error handling. The errors your database give you for schema violations aren’t really user friendly, so I have to convert them to proper errors (you must not enter negative amounts). When using SQLite there didn’t seem a way to know what column cast the error without parsing the string. That often lead that I had to implement the business logic twice. One check for in the code for the error handling and another one in the database. That was also often quite inefficient and prone to race conditions (if using improper isolation). For example if the table has two unique fields, I cannot just check for a unique constraint violation because it doesn’t tell me the column in the error data structure.

Looking at the Postgres driver it seems easier but is there any good tutorial on how to do error handling properly for databases?