sorting rails view with will_paginate plugin

This solution for sorting data on rails index actions does NOT use ajax. Sometimes sorting with ajax is nice but often it is not actually as user friendly; breaking the browser back button, etc...

Disclaimer 1: this code is probably using all kinds of ruby anti-patterns :) - but it is a very simple implementation and so I just went with it...

Disclaimer 2: the title of the post might be a little miss leading as you can use this with or without the will_paginate plugin. In my case I am using it with the will_paginate plugin - and I highly recommend it!

put this file in your rails lib directory /lib/sort_index.rb

	# see http://gist.github.com/341290 (sort_index_controller_usage.rb)
	# see http://gist.github.com/341295 (sort_index_view_usage.html.erb)
	# The classes in this module help to enable sorting on index pages
	# building sql order clauses and rendering html table header links
	module SortIndex
	SORT_KEY_ASC = 'asc'
	SORT_KEY_DESC = 'desc'
	SORT_DIRECTION_MAP = {
	SORT_KEY_DESC => 'DESC',
	SORT_KEY_ASC => 'ASC'
	}
	# The +SortIndex::Config+ class specifies all possible sort columns
	# for a given controller action including the default column and the default order
	class Config
	attr_accessor :columns
	attr_accessor :default_direction
	def default
	return @default
	end
	# The +initialize+ method;
	# both the default and columns parameters contain key value pairs
	# where the key is passed in the query string to the action and the
	# value contains the sql order by value
	# === Parameters
	# * _default_ = Hash; must contain only one pair; automatically gets added to the columns member
	# * _columns_ = Hash; one pair per sortable column
	# * _default_direction_ = String; optional, if not specified order will be DESC
	def initialize(default, columns, default_direction = nil)
	@columns = columns
	@default_direction = default_direction || SORT_KEY_DESC
	raise "default only supports 1 pair" if default.length != 1
	default.each_pair { |key, value|
	@default = value
	@columns[key] = value
	}
	end
	end
	# The +SortIndex::Sortable+ class enable sorting on index pages
	# avoids sql injection by only using values from the SortIndex::Config#columns
	# Hash and not the values passed in the query string
	class Sortable
	# The +initialize+ method;
	# === Parameters
	# * _params_ = the controllers params Hash
	# * _config_ = SortIndex::Config
	# * _index_url_ = String; optional, if not specified will be the name of the controller
	# ** Examples
	# *** not specified /employees (the index action)
	# *** specified /employees/special_action
	def initialize(params, config, index_url = nil)
	@config = config
	@params = params
	@index_url = index_url || params[:controller]
	# sets up for building the sql order by
	@sort_direction = SORT_DIRECTION_MAP[@params[:sort_direction]] || @config.default_direction
	@sort_by = @config.columns[@params[:sort_by]] || @config.default
	end
	# The +order+ method returns the sql order criteria
	# use with your find calls or via paginate from will_paginate plugin
	def order
	specified_sort_by || "#{@sort_by} #{@sort_direction}"
	end
	# The +header_link+ method returns a string of html containing the table header and a tags
	# Example: <th><a href="/employess?sort_by=first_name&amp;sort_direction=desc">First Name</a></th>
	# If the column is the currently sorted column then a css class of current-sort-asc or current-sort-describe
	# is applied; this allows you to use css to add visual indicators such as up and down arrows
	# this method is called from the view; once per column
	# === Parameters
	# * _sort_key_ = String; must be one of the key values from SortIndex::Config
	# * _display_ = The display text
	# * _sortable_ = Boolean; default is true;
	# ** passing false will not render an anchor tag; instead the display will be wrapped in a span
	def header_link(sort_key, display, sortable = true)
	if @config.columns[sort_key].nil? and sortable then
	raise "Sort key of '#{sort_key}' not found. Check your controllers SortIndex::Config variable"
	end
	class_attr = ""
	if @config.columns[sort_key] == @sort_by then
	class_attr = " class='current-sort-#{@sort_direction.downcase}'"
	end
	a_href = "<a href=\"#{@index_url}?sort_by=#{sort_key}&amp;sort_direction=#{next_direction}\" title=\"Sort by #{display}\">#{display}</a>"
	if sortable == false then
	a_href = "<span>#{display}</span>"
	end
	return "<th#{class_attr}>#{a_href}</th>"
	end
	# The +next_direction+ method is called by header_link and specifies which way the rendered
	# links should sort. Returns the opposite of the current sort
	def next_direction
	sort_direction = SORT_KEY_ASC
	if (@params[:sort_direction].nil?) then
	sort_direction = (@sort_direction == SORT_KEY_ASC) ? SORT_KEY_DESC : SORT_KEY_ASC
	elsif (@params[:sort_direction] == SORT_KEY_ASC) then
	sort_direction = SORT_KEY_DESC
	end
	return sort_direction
	end
	# The +specified_sort_by+ method is called by order returns the sql order by criteria
	# This can be more than one sql column; when it is multiple columns we apply the same direction to all
	# For Example if you had one header column for Employee#full_name which mapped to two
	# database columns first_name and last_name of the employees table the result would look like
	# first_name DESC, last_name DESC or last_name DESC, first_name DESC depending on your configuration
	def specified_sort_by
	sort = @config.columns[@params[:sort_by]]
	return nil if sort.nil?
	return sort.split(',').map {|order_criteria| "#{order_criteria} #{@sort_direction}"}.join(',')
	end
	end
	end

view raw sort_index.rb hosted with ❤ by GitHub

in your controller code, set up the SortIndex::Config (you can have more than one if you have multiple actions that need to support sorting

	# see http://gist.github.com/341278 (sort_index.rb)
	# see http://gist.github.com/341295 (sort_index_view_usage.html.erb)
	# this example is an employees controller and using the will_paginate plugin
	# however code should work fine with a standard ActiveRecord#find(:all, :order => ...
	# put the sort_index.rb code in your rails lib directory
	class EmployeesController < ApplicationController
	# index sort constant config for sorting index action
	# default is order by updated_at DESC
	INDEX_SORT = SortIndex::Config.new(
	{'updated_at' => 'updated_at'},
	{
	'full_name' => 'UPPER(first_name), UPPER(last_name)',
	'email' => 'email'
	}
	#, optionally SortIndex::SORT_KEY_ASC
	)
	def index
	@sortable = SortIndex::Sortable.new(params, INDEX_SORT)
	@employees = Employee.paginate :page => params[:page], :order => @sortable.order
	# render index.html.erb
	end
	end

view raw sort_index_controller_usage.rb hosted with ❤ by GitHub

then in your view code render your table headers using the sort object

	<%-
	# see http://gist.github.com/341278 (sort_index.rb)
	# see http://gist.github.com/341290 (sort_index_controller_usage.rb)
	-%>
	<table style="width:100%">
	<thead>
	<tr>
	<%= @sortable.header_link('full_name', 'Name') %>
	<%= @sortable.header_link('email', 'Email') %>
	<%= @sortable.header_link('updated_at', 'Updated at') %>
	<%= @sortable.header_link('manage', 'Manage', false) %>
	</tr>
	</thead>
	<tfoot>
	<tr>
	<td colspan="4">
	&nbsp;
	</td>
	</tr>
	</tfoot>
	<tbody>
	<% for employee in @employees %>
	<tr class="<%= cycle('oddrow', 'evenrow') %>">
	<td><%= employee.full_name %></td>
	<td><%= employee.email %></td>
	<td><%= employee.updated_at.to_s %></td>
	<td>
	<%= link_to 'Edit', edit_employee_path(employee) %>
	</td>
	</tr>
	<% end %>
	</tbody>
	</table>

view raw sort_index_view_usage.html.erb hosted with ❤ by GitHub

Does not support the following:

• additional attributes on the anchor tags
• additional attributes on the table header tags
• additional query string parameters - might add this later, would be nice for search results

Ruby Http Get with Net::HTTP

	require 'net/http'
	require 'uri'
	def get_html_content(requested_url)
	url = URI.parse(requested_url)
	full_path = (url.query.blank?) ? url.path : "#{url.path}?#{url.query}"
	the_request = Net::HTTP::Get.new(full_path)
	the_response = Net::HTTP.start(url.host, url.port) { |http|
	http.request(the_request)
	}
	raise "Response was not 200, response was #{the_response.code}" if the_response.code != "200"
	return the_response.body
	end
	# this will fail with ArgumentError: HTTP request path is empty
	s = get_html_content("http://www.google.com")
	# these should be fine
	s = get_html_content("http://www.google.com/")
	s = get_html_content("http://github.com/search?q=http")
	# above code does not handle redirects but raises exception for non-200
	s = get_html_content("http://www.yahoo.com/") # http 302

view raw Net::HTTP http get.rb hosted with ❤ by GitHub

Resources

• Net::HTTP ruby docs
• Stackoverflow post

require 'net/http'
require 'uri'

def get_html_content(requested_url)
url = URI.parse(requested_url)
full_path = (url.query.blank?) ? url.path : "#{url.path}?#{url.query}"
the_request = Net::HTTP::Get.new(full_path)

the_response = Net::HTTP.start(url.host, url.port) { |http|
http.request(the_request)
}

raise "Response was not 200, response was #{the_response.code}" if the_response.code != "200"
return the_response.body
end

# this will fail with ArgumentError: HTTP request path is empty
s = get_html_content("http://www.google.com")
# these should be fine
s = get_html_content("http://www.google.com/")
s = get_html_content("http://github.com/search?q=http")
# above code does not handle redirects but raises exception for non-200
s = get_html_content("http://www.yahoo.com/") # http 302

Ruby Regex to remove script tags

	# remove all script tags
	html_content = html_content.gsub(/<script.*?>[\s\S]*<\/script>/i, "")

view raw ruby regex remove html script tags.rb hosted with ❤ by GitHub

Resources

• Rubular
• Match any character in RegEx
• Ruby String gsub