I’m writing a distributed system simulator in Ruby and had a need to verify that an array contained only integer process ID’s, and no elements of any other type. Using Ruby’s Enumerable#inject function, this is a wonderfully quick one-liner.

>> a = [1,2,3]
=> [1, 2, 3]
>> a.inject(true) do |result, element| result && element.is_a?(Fixnum) end
=> true
>> b = [1,"Hi",3]
=> [1, "Hi", 3]
>> b.inject(true) do |result, element| result && element.is_a?(Fixnum) end
=> false

Edit: After I posted this, Robert Riemann pointed out that Enumerable#inject iterates over the entire array even if the first value in the array is not an integer. This is easily confirmed as follows:

>> a.inject(true) do |result,element| puts "Next element"; result && element.is_a?(Fixnum) end
Next element
Next element
Next element
=> false

So, inject is not what we’re looking for in this case. Later, Martin Vidner offered a solution using Enumerable#all?:

>> a = [1,2,3]
=> [1, 2, 3]
>> a.all? do |e| e.is_a?(Fixnum) end
=> true
>> b = [1,"Hi",3]
=> [1, "Hi", 3]
>> b.all? do |e| e.is_a?(Fixnum) end
=> false

Using all?, we see that processing stops after a non-integral value is encountered:

>> b = [1,"Hi",3]
=> [1, "Hi", 3]
>> b.all? do |e| puts "Next element"; e.is_a?(Fixnum) end
Next element
Next element

Thanks to both Robert and Martin for pointing out the problem in my original solution.

As part of my Christmas reading, I am currently devouring the fantastic Ruby Best Practices from O’Reilly, which is packed with a plethora of tricks that remind you on each page why Ruby programming is joyous and beautiful.

The book pointed out one particular trick of which I wasn’t aware. When you want to capture a regular expression match from a string, you can simply index into the string with the regular expression, and the capture number.

>> str = "First name: Jeff      Last name: Shantz"
=> "First name: Jeff      Last name: Shantz"
>> first_name = str[/:\s*(\w+)[^:]+:\s*(\w+)/,1]
=> "Jeff"
>> last_name = str[/:\s*(\w+)[^:]+:\s*(\w+)/,2]
=> "Shantz"