similar to: File.open behavior for ActiveRecord, to ensure save after manipulation?

Hello, I am attempting to create several plots based on "site" (~300 total) and am having trouble with the code. I simply want to create a plot using the code, plot(year, peak), for the following dataset. I would like for each site to be plotted on a separate page and the plots saved in a directory. Would a "foreach" loop work? I tried a "by" statement, but

Hi I was recently taught about the useful ActiveRecord::Base#serialize. It''s very useful for persisting arbitrary objects through the regular database. For example, say you want to store a Set of all the years a Convention took place. You do: class Convention < ActiveRecord::Base serialize :years, Set end This effectively creates a Convention#years attribute, which is a set: you

So I want to SELECT from multiple tables, so as to be able to do "SELECT FROM foo, bar WHERE foo.id=bar.id AND bar.baz=23". How do I do it with AR#find? Also, I need this for associations. has_many et. al. have :finder_sql, which is nice, but feels like a Pyrrhic: it''s basically doing the association in pure SQL, defeating the purpose of AR. -- -Alder

Hi there I''m planning to build "The Perfect Rails/Debian/Lighttpd Stack", published by Ezra Zygmuntowicz and Sean Schertell at http://brainspl.at/rails_stack.html I''m wondering what sort of constant maintanance that setup would incur. What actions have be regularly performed to keep this setup healthy and secure under normal conditions? Alder

Hi Let''s say we set session[:foo] = ''bar''. Now we want to remove the :foo key completely. How do you do that? Appearantly, session is a CGI::Session [1] instance, so it doesn''t have #delete like Hash. I tried: session[:foo] = nil But then debug(session) shows an empty :foo key. I''d like to remove that key, as my session is already pretty crowded. Is

Hi there I need to store additional attributes on join table. Searching the Wiki [1] revealed two possible methods: 1) push_with_attributes, as described by Justin Palmer [2]. Unfortunately, "this method is now deprecated" [3]. Even if it wasn''t, there seem to be inherent problems with that approach, for example with updating those additional attributes, though Joshua Muheim

Hi I have an action which saves the originating resource from which it was called, and returns to it later. This is done by setting: session[:return_to] = request.referer ... redirect_to(session[:return_to]) Which works great for GET resources. The problem is that POST variables aren''t kept in request.referer, so this method fails when the originating resource is a a POST.

Hi, for some models I found that it''s easier to share the controller new and update code just to share the view code. At the front of every such action function I have to lookup if the object already exists, if so "find" it from the database and update the "attributes=". If there is no such object, "new" it with the attributes I''ve got from the

Hi I''m choosing a database for a new Rails application. I pretty much ruled out PostgreSQL, since despite good reviews of the database itself, it seems it has countless issues with Rails (for example: in recent releases migrations don''t work, and unit-tests fail). So I should now choose between MySQL 4.1, which as I understand enjoys widespread popularity - perhaps the most

Hi I created a custom route that looks like this: map.connect '':controller/:table/:action/:id'' Hierarchically, :table is above :action and :id. So, if we call ActionController::Base#url_for with an argument :table=>nil, we''re supposed to get a url with :action and :id set to nil as well. However, this is not the case: url_for(:table=>nil) from

Hi there I''m considering Rails for a new web-application planned to be built from scratch. The application has an extensive web front, and in fact most of the application is interfaced through that front and engaged through clients'' HTTP requests (i.e. the conventional way web-applications and "dynamic web-sites" work). However, some essential parts are supposed to

A concrete example: We''re building a system to organize the information about workshops being held in various conventions. A convention has more than one workshop, and each workshop can be held in more than one convention. Also each workshop has a host, who is specific to a workshop being held in a specific convention. For example, Matz may host an "Introduction to Ruby"

Hi Let''s say we have model Foo. Each Foo instance can have several bars. Those bars are primitive, so they shouldn''t be models. For example, Foo might be a type of convention, and the bars might be years the convention was held in. Naively, we would have a conventions_years date, and put: has_many :years inside class Convention. But then we''d get an error, since for

I''m looking to implement model inheritance in a new application. Is there any good alternative to single-table-inheritance? -- -Alder

Hi ActionView::Helpers::FormHelper methods (e.g. #text_field, #check_box...) take object_name as an argument. From this argument they infer a name that is supposed to contains the reference to the relevant object. I can''t see why the FormHelper methods were designed that way. Compared to simply passing them a reference, the current way seems too complicated and an invitation to problems

Hi everyone. I''m choosing an SCM for my current Rails project. Up until now, for my other (non-Rails) codings, I used Darcs and liked it very much. I''d prefer to use Darcs for my Rails development as well. Looking into the matter, I''m somewhat deterred by the appearant dominance of Subversion as the SCM tool of choice for Rails hackers. Is there a technical reason for

I have an ActiveRecord class class User < ActiveRecord::Base end and I want to define an attribute, in example "links", links will be an array. class User < ActiveRecord::Base attr :links end and now, I want links to be an array, so I do class User < ActiveRecord::base attr:links def initialize(attributes) super(attributes) @links =

Hello, I need to access an mirrored database with one read and one write cluster. Is there a way to tell a model to use two connections? So that all updates etc go to the write database and all read from the read one? This is a fixed set up I can not use MySQL 5.x stuff and additionally i get a token on write, which can be checked on read (and wait for replication) if new data is needed. I

Recently I wanted to check that my test data was valid. Seems like a reasonable thing to want to do so I thought I''d share the rake task. Add this code to a rake file in lib/tasks and run with rake db:fixtures:validate. It will print out the class, record id and error messages for any model objects that fail validation. Enjoy! -Jonathan. -- namespace :db do namespace :fixtures do

Hi. When you have library foo.rb in your Rails project /lib directory, I notice that on some platforms - e.g. Windows with WEBrick as a server - you can require it by simply: require ''lib/foo'' However on other platforms - e.g. Debian with Lighttpd server via FastCGI - this does not work, and you need to specify RAILS_ROOT explicitly, like this: require