You're right that it's a "dev/dba problem", but the reason for not using `select !<column>` is exactly the same as not using `select *`. If you specify exactly what you want, there's less of a chance of your ORM-based app breaking because it received an extra or too few columns. I'm sure most people who work with SQL DBs have been bitten this before.
Also, while it could also be useful, over the years I've started moving away from things like using `!` to negate things. It makes it slightly harder to read and it ruins the symmetry of my code. And lord knows SQL doesn't need any more of that. It's probably the only language I use often where I have to place my commas at the start of the line; where I need to add a first filter as `1 = 1` etc just to make it easier to comment out things quickly.
What I personally want most is the ability to move the FROM clause to the top (this is a no-brainer) and better aliasing rules. The alias I used in my SELECT should be accessible everywhere in the statement.
Also, while it could also be useful, over the years I've started moving away from things like using `!` to negate things. It makes it slightly harder to read and it ruins the symmetry of my code. And lord knows SQL doesn't need any more of that. It's probably the only language I use often where I have to place my commas at the start of the line; where I need to add a first filter as `1 = 1` etc just to make it easier to comment out things quickly.
What I personally want most is the ability to move the FROM clause to the top (this is a no-brainer) and better aliasing rules. The alias I used in my SELECT should be accessible everywhere in the statement.