Mark One and Only One Record as "Primary" - mysql

Question

I'm writing a genealogy application and I'm trying to figure out on aspect of my database model.
I have a table for people:
create table person (
id int unsigned not null primary key
)
And I have a table for names:
create table name (
id int unsigned not null primary key,
person_id int unsigned not null,
first_name varchar(191),
last_name varchar(191),
foreign key (person_id) references person (id)
)
A person can have multiple names because sometimes they have different names in different censuses and other records.
But a person also has one particular name which we should display by default.
My first thought was to add a primary column to the table name:
create table person (
id int unsigned not null primary key
)
create table name (
id int unsigned not null primary key,
person_id int unsigned not null,
first_name varchar(191),
last_name varchar(191),
`primary` tinyint(1) not null default 0,
foreign key (person_id) references person (id)
)
This has a data-integrity issue. Multiple names that belong to a single person could have the primary flag set.
Another way I thought of was to add a primary_name_id column to the person table. This could link back to the primary name in the name table:
create table person (
id int unsigned not null primary key,
primary_name_id int unsigned default null,
foreign key (primary_name_id) references name (id)
)
create table name (
id int unsigned not null primary key,
person_id int unsigned not null,
first_name varchar(191),
last_name varchar(191),
foreign key (person_id) references person (id)
)
This also has has an integrity issue. primary_name_id might point to a name row that belongs to a different person. Also, it requires the primary_name_id column to be nullable because when a person is first created he has no names yet.
The third way I thought about doing this is to include duplicates of the "name" fields in the person record:
create table person (
id int unsigned not null primary key,
first_name varchar(191),
last_name varchar(191),
)
create table additional_name (
id int unsigned not null primary key,
person_id int unsigned not null,
first_name varchar(191),
last_name varchar(191),
foreign key (person_id) references person (id)
)
This doesn't seem like an ideal solution either. Similar kinds of data are being stored in two places. Also changing a person's primary name would require a bit of work. I'd have to insert a new additional_name, update the person, and delete an old additional name.
Is there a better way to do this?

Answer 1

Your second method is basically the right way. But, you want to be sure that the name is for the person. So:
create table person (
id int unsigned not null primary key,
primary_name_id int unsigned default null,
foreign key (id, primary_name_id) references name (person_id, id)
);
create table name (
id int unsigned not null primary key,
person_id int unsigned not null,
first_name varchar(191),
last_name varchar(191),
foreign key (person_id) references person (id),
unique (person_id, id);
);
The unique constraint is a bit of redundancy, but it let's you ensure that the values match across the two tables -- appropriately.