The constraint types in PostgreSQL

• NOT NULL — a value is required.
• CHECK — a boolean condition each row must satisfy.
• UNIQUE — no two rows share the value(s).
• PRIMARY KEY — UNIQUE + NOT NULL; the row identity.
• FOREIGN KEY (REFERENCES) — the value must exist in another table, with an action on delete/update.
• EXCLUDE — a generalised uniqueness (e.g. no overlapping bookings), PostgreSQL-specific.
• DEFAULT is not strictly a constraint but is declared alongside them.

create table enrolments (
    id          bigint generated always as identity primary key,
    student_id  bigint not null references students(id) on delete cascade,
    course_code text   not null,
    marks       int    check (marks between 0 and 100),
    enrolled_on date   not null default current_date,
    unique (student_id, course_code)
);

Constraints keep bad data out no matter which application writes to the database, which is why they are better than validation in application code alone. Name them (constraint marks_range check (...)) to get clearer error messages.