In my book the only thing a better sql needs is to start selects with the from clause. -‘from orders select orderid’ Which would make for better autocomplete among other things.
In the last 10 years of my career, many a dev & prod databases have been accidentally harmed by an overzealous query.
This could also add a warning for "left-over" non-updated rows, for example if the developer signals that only one row should be updated, but the WHERE clause has many rows, its usually a sign of a missing unique key, something I've seen before as well.
So you change:
1. What I want
2. Which tables
3. Which filter
4. What grouping
Putting what you want further down the list might be more confusing rather than less. (a + b) where
a = … * c,
b = … * sqrt(2) + d
c = …,
d = … …;
But that is not less confusing, that is more. I need eggs, flour, sugar and vanilla from the supermarket.
Vs Yoda From the supermarket eggs, flour, sugar and vanilla I need.
The original use case has long failed, but here we are with SQL stockholm syndrome. I don't hate it, but strictly speaking it is a bit odd.Edit: there was a linguist in the group that developed SQL to help with the goal of making it easier for non programmers.
[0] https://www.mcjones.org/System_R/SQL_Reunion_95/sqlr95-Syste...