From: Phrogz <phrogz@...>
Date: 2010-04-20T11:15:05+09:00
Subject: Re: Elegant Solution to a Seemingly Simple Problem?

On Apr 18, 9:15 pm, Derek Cannon <novelltermina...@gmail.com> wrote:
> >doc = Nokogiri::HTML(open(url))
> >   table = []
> >   doc.css("tr").each do |row|
> >     cells = row.css("td").map {|cell| cell.text.strip }
> >     next unless cells.size == 4
> >     next unless cells[1] == "TBA"
> >     cells.insert(2, "")
> >     table << cells
> >   end
>
> This is interesting. So nil is returns for elements that activate the
> unless statement?

No. Next inside a block is like an early return. The ruby code:
   do_this unless something
is the same as:
   unless something
     do_this
   end
is the same as:
   if !something
     do_this
   end

So what's happening in the above is that you're creating an array
(table), and then looping through the list of rows returned from
Nokogiri. For each row, if certain criteria are met (there aren
exactly four cells, or the second cell is "TBA") you immediately stop
processing that row and move on to the next. ("Next!" shouts the
government worker, dismissing you and moving on.) If you didn't bail
out early, however, you inject an extra entry into the array of
strings and then shove that whole array as a new entry onto the end of
the table array.

Because you're using the pattern of creating an array and
conditionally populating during a traversal, there are no embedded
nils to clean up later.

For comparison, here's similar code that WOULD leave you with a table
with some nil entries (that you'd probably want to #compact
afterwards):

  table = doc.css("tr").map do |row|
    cells = row.css("td").map {|cell| cell.text.strip }
    if cells.size==4 && cells[1]=="TBA"
      cells.insert(2, "")
      cells
    end
  end

In the code immediately above, the value of an if statement that
matches is the value of the last expression, while the value of an if
statement that is not matches is nil.

Here's a simpler example showing the difference between map and each-
with-injecting:

irb(main):001:0> digits = (1..9).to_a
=> [1, 2, 3, 4, 5, 6, 7, 8, 9]

irb(main):002:0> odds = []
=> []

irb(main):003:0> digits.each{ |n| if n%2==1 then odds<<n end }
=> [1, 2, 3, 4, 5, 6, 7, 8, 9]

irb(main):004:0> odds
=> [1, 3, 5, 7, 9]

irb(main):005:0> evens = digits.map{ |n| if n%2==0 then n end }
=> [nil, 2, nil, 4, nil, 6, nil, 8, nil]

irb(main):006:0> evens.compact
=> [2, 4, 6, 8]