While creating a schema from a database many-to-many relationships between tables are not created.
Is this a principal problem?
Is it possible to detect from the table structure that many-to-many relationships exist and create the respective code in schema classes automagically?
The problem with developing this kind of code is that many tables that contain multiple references are not many-to-many tables, and have multiple references for other reasons. For instance, I'll make up a schema for some fictional app where something could be regarded as a many-to-many table, when it is not.
create table category (
id primary key,
...
);
create table sub_category (
id primary key,
category references category(id),
...
);
/* EDIT:
This is the table that could be regarded as many_to_many
by an automated system */
create table product (
id primary key,
category references category(id),
sub_category references sub_category(id),
...
);
Something could be built this way for ease of use, without having to do multiple table joins in the database on a website, especially when considering speed. It would be difficult for a piece of code to say definitively 'this is not a many_to_many' situation, while the developer should be able to easily figure it out, and add in the many_to_many line below the checksum.
I consider DBIX::Class schema outputs a good starting point, and little more, especially when working with auto numbering in non-MySQL databases, among other things. I often need to modify above the "Don't modify above this line" stuff (although many_to_many can obviously go below that checksum, of course.