My setup is the following: I am using Rails 3.0.1 with a PostgreSQL 9.0 database. I have my models organized in submodules.
When I'm using the console to create a new object of any kind (c = ::Physical::Base::Contact.new
) and save it, the id of the object is not automatically filled in. It is written to the database, but my object that I have just created does not take note of the ID that is generated by the sequence in PostgreSQL. My migration script for creating the table is below. I write them myself and am not using the standard rails generation feature to have more control over my tables.
Is this a standard PostgreSQL+Rails behaviour, or am I missing something by writing the migrations myself?
class Contact < ActiveRecord开发者_如何学C::Migration
def self.up
execute "
create table contacts(
id serial,
contact_type_id integer not null check (contact_type_id > 0),
value text not null check (length(value) > 0)
)
"
end
def self.down
execute "
drop table contacts;
"
end
end
Make sure you set a primary key on your table, I had the same problem recently, and it was very frustrating. I decided to re-create the table using migrations and noticed my old table was absent a primary key. Once I added it, things worked as expected!
You'll need something like "alter table contacts add primary key(id)" -- ActiveRecord doesn't seem to understand if there's not a primary key index on that field.
Before the primary key I would get:
ruby-1.9.2-p0 > Job.create
=> #<Job id: nil, data: nil, timestamp: nil, enqueued_at: nil, started_at: nil, completed_at: nil, status: nil, result: nil, job_queue_id: nil, job_handle: nil, uniq: nil, priority: nil, server_id: nil>
Now that I've added the primary key, the result is:
ruby-1.9.2-p0 > Job.create
=> #<Job id: 17937, data: nil, timestamp: nil, enqueued_at: nil, started_at: nil, completed_at: nil, status: nil, result: nil, job_queue_id: nil, job_handle: nil, uniq: nil, priority: nil, server_id: nil>
精彩评论