I just started working at a software company that has never had a DBA. The database that evolved over time has 1157 Tables with over 28000 columns. Some table are small but many have 200+ columns.The Problem: All of the SQL is built dynamically in the application and the SQL in Most cases looks like 'SELECT t1.*, t2.* from ........'. Even when the application only uses a couple of the columns returned, the SQL does a SELECT *.I have spoken to the lead developer and told him that it is a very bad practice and his response is always, 'Prove It'. So, I am looking for good ways to Prove that doing a SELECT * from a 200 column table when you only need 2 of the columns is a Very bad practice. I want to show things like performance, resource utilization, caching issues, and anything else that you bright professionals could turn me on to. Descriptions of how to do it are good, scripts that show it are even better.Thanks in advance!Mark.
↧