如何跳过Ruby哈希中的空值

Question

I have an array in which each array item is a hash with date values, as shown in my example below. In actuality, it is longer and there are about 20 dates per item instead of 3. What I need to do is get the date interval values for each item (that is, how many days between each date value), and their intervals' medians. My code is as follows:

require 'csv'
require 'date'

dateArray = [{:date_one => "May 1", :date_two =>"May 5", :date_three => " "}, {:date_one => "May 10", :date_two =>"May 10", :date_three => "May 20"}, {:date_one => "May 6", :date_two =>"May 11", :date_three => "May 12"}]

public
def median
sorted = self.sort
  len = sorted.length
  return (sorted[(len - 1) / 2] + sorted[len / 2]) / 2.0
end

puts dateIntervals = dateArray.map{|h| (DateTime.parse(h[:date_two]) - DateTime.parse(h[:date_one])).to_i}
puts "\nMedian: " 
puts dateIntervals.median

Which returns these date interval values and this median:

4
0
5
Median: 4

However, some of these items' values are empty, as in the first item, in its :date_three value. If I try to run the same equations for the :date_three to :date_two values , as follows, it will throw an error because the last :date_three value is empty.

It's okay that I can't get that interval, but I would still would need the next two items date intervals (which would be 10 and 1).

How can I skip over intervals that return errors when I try to run them?

Answer 1

I would recommend adding helper functions that can deal with the types of inputs you're expecting. For instance:

def date_diff(date_one, date_two)
    return nil if date_one.nil? || date_two.nil?
    (date_one - date_two).to_i
end

def str_to_date(input_string)
    DateTime.parse(input_string)
    rescue
    nil
end

dateArray.map{|h| date_diff(str_to_date(h[:date_three]), str_to_date(h[:date_two])) }
=> [nil, 10, 1]

dateArray.map{|h| date_diff(str_to_date(h[:date_three]), str_to_date(h[:date_two])) }.compact.median
=> 5.5

The bonus here is that you can then add unit tests for the individual components so that you can easily test edge cases (nil dates, empty string dates, etc).

Answer 2

In your map block, you can just add a check to make sure the values aren't blank

dateIntervals = dateArray.map{ |h| 
  (DateTime.parse(h[:date_two]) - DateTime.parse(h[:date_one])).to_i unless any_blank?(h)
}

def any_blank?(h)
  h.each do |k, v|
    return true if v == " "
  end
end

Answer 3

I would first just filter out the empty values first (I check if the string consists entirely of whitespace or is empty), then compare the remaining values using your existing code. I added a loop which will compare all values in the sequence to the next value.

dateArray = [
  { date_one: "May 1", date_two: "May 5", date_three: " ", date_four: "" },
  { date_one: "May 10", date_two: "May 10", date_three: "May 20" }
]

intervals = dateArray.map do |hash|
  filtered = hash.values.reject { |str| str =~ /^\s*$/ }
  (0...filtered.size-1).map { |idx| (DateTime.parse(filtered[idx+1]) - DateTime.parse(filtered[idx])).to_i }
end

# => [[4], [0, 10]]