# Seed\_fu and has\_many :through **URL:** \ **Category:** Rails\ **Created:** [March 29, 2010, 2:22pm UTC](https://www.ruby-forum.com/t/seed-fu-and-has-many-through/185938 "2010-03-29T14:22:43Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![tom-mac](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/t/bc8723/32.png) [@tom-mac](https://www.ruby-forum.com/u/tom-mac)\ **Post date:** [March 29, 2010, 2:22pm UTC](https://www.ruby-forum.com/t/seed-fu-and-has-many-through/185938/1 "2010-03-29T14:22:43Z")
Hi In my application to seed data I am using seed\_fu ([GitHub - mbleigh/seed-fu: Advanced seed data handling for Rails, combining the best practices of several methods together.](http://github.com/mbleigh/seed-fu)) ``` I have tables roles,privileges and permissions( It joins roles and ``` privileges). Now at application startup I have to seed these tables with data. But during runtime some additionals rows may be added to permissions. But I dont how at a later stage if I run rake db:seed\_fu all data that I set later to permissions may be wipeout. I dont know how to handle this. As an example # role 1 admin 2 contact 3 employee # privilege 1 add\_contact 2 edit\_contact 3 delete\_employee # permissions (role\_id,privilege\_id) 1 1 1 2 3 2 …etc ``` Suppose this may be seeded at startup. Later additional entries ``` may go to permissions. One of my question is how can I use seed\_fu to fill data for permissions table? Can I do it as usual like create db/fixtures/permission.rb and Permission.seed\_many(:id, :id, [ { :role\_id =\> 1, :privilege\_id =\> 1 }, { :role\_id =\> 1, :privilege\_id =\> 2 }, { :role\_id =\> 3, :privilege\_id =\> 2 } ]) ``` One more question is if I follow like above and at a later stage if ``` run rake db:seed\_fu again all data that where added later to permission will be lost. And also how can I ensure that all the id values above should be what I expected. Please guide me Thanks in advance Tom ---
**Author:** ![tom-mac](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/t/bc8723/32.png) [@tom-mac](https://www.ruby-forum.com/u/tom-mac)\ **Post date:** [March 29, 2010, 3:28pm UTC](https://www.ruby-forum.com/t/seed-fu-and-has-many-through/185938/2 "2010-03-29T15:28:19Z")
On 29 March 2010 13:22, Tom M. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > Hi > In my application to seed data I am using seed\_fu > ([GitHub - mbleigh/seed-fu: Advanced seed data handling for Rails, combining the best practices of several methods together.](http://github.com/mbleigh/seed-fu)) > > I have tables roles,privileges and permissions( It joins roles and > privileges). Now at application startup I have to seed these tables with > data. But during runtime some additionals rows may be added to > permissions. But I dont how at a later stage if I run rake db:seed\_fu > all data that I set later to permissions may be wipeout. I dont know how > to handle this. This is presumably a follow on to your previous thread. You never explained why you are repeatedly seeding the database, this is normally something that would be done only once. Colin ---
**Author:** ![tom-mac](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/t/bc8723/32.png) [@tom-mac](https://www.ruby-forum.com/u/tom-mac)\ **Post date:** [March 30, 2010, 6:05am UTC](https://www.ruby-forum.com/t/seed-fu-and-has-many-through/185938/3 "2010-03-30T06:05:16Z")
Hi Colin ``` Suppose I have the development phases like phase1,phase2, etc. So ``` for example in Phase1 I have to seed data to say table1 (rake db:seed\_fu) And it is deployed. Now again in phase2 again I have to seed data to table2 . Then also to sseed data I run rake db:seed\_fu ``` But when ever I run this command it seeds data to table1 (once ``` more) and table2 (for the first time). Is n’t it? I think I can answer your question. But could you please give answer to my original questions posted in this thread. Thanks Tom ---
**Author:** ![tom-mac](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/t/bc8723/32.png) [@tom-mac](https://www.ruby-forum.com/u/tom-mac)\ **Post date:** [March 30, 2010, 9:44am UTC](https://www.ruby-forum.com/t/seed-fu-and-has-many-through/185938/4 "2010-03-30T09:44:35Z")
On 30 March 2010 05:05, Tom M. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > more) and table2 (for the first time). Is n’t it? I think I can answer > your question. But could you please give answer to my original questions > posted in this thread. Remove the data for the first table from the seed code once you have done it, so that the next time it just seeds the second one. You can always get the old version back from your version control system if you ever need it again. Colin