Learning SQL is easy; Implementing SQL, not so much. Halloween has come and gone. But, the tricks of the Halloween problem is here to stay! This has to be solved by databases every day. SQL made the relational database easy, accessible, and successful. SQL may be easy to write, but its implementation hides a lot of complexity. From the days of System R to NoSQL systems, all of us implementing SQL and SQL inspired languages like N1QL , have to learn the rules, requirements to ensure the declarative query language continues its extraordinary effectiveness and correctness. One such problem is the Halloween Problem . I'll describe the problem briefly here. I highly recommend reading its history in the IEEE Annals of the history of computing - second link in the references section below. The Problem: Consider a simple table and index: Java 1 CREATE TABLE t ( empid int primary key , salary int ); 2 CREATE INDEX i1 ON t ( salary...
A blog about all things data and data processing issues and interests. SQL, NoSQL, flexible schema, scale-up, scale-out, transactions, high availability.