I have a feeling that the fact that SQL is inherently unordered is very important for making it fast and distributed. I am thinking about map/reduce types of j

I have a feeling that the fact that SQL is inherently unordered is very important for making it fast and distributed. I am thinking about map/reduce types of jobs. Is that something you have solved?