Submitted by: Juarez Rudsatz (juarezr)

Columnar storage[1] is the preferred format for OLAP, Business Intelligence and Big Data processing storage and consumption.
It is credited to perform faster processing queries on tables following the Star schema[2] used to develop data warehouses.

The syntax could follow the actually used, plus some specific details of columnar tables like column encoding, table sort order/clustering.

Referential integrity could be optional because of people usually loads/updates big sets of data and primary/foreign key checking becomes a burden. Primary/foreign key checking is util in ddl usage when creating/dropping tables/fields.

Mixing regular row based and columnar tables in queries and database storage could be also allowed, but usually, they are separated use cases.

Of course, this is not a trivial enhancement because it impacts top down in the database engine and adds a second path for developing/maintaining all query/storage algorithms and code.

[1] https://en.wikipedia.org/wiki/Column-oriented_DBMS
[2] https://en.wikipedia.org/wiki/Star_schema