Ruby: Iterate over entire array from nth position - ruby

I'd like to iterate over an entire array, starting from any position. I'm not sure if there's a way to achieve this easily in Ruby, and I couldn't find any examples in the Array or Enumerator docs.
array = [0, 1, 2, 3, 4]
array.each.starting_at(3) { |e| e }
#=> [3, 4, 0, 1, 2]
And also:
array.each.starting_at_reverse(3) { |e| e }
#=> [3, 2, 1, 0, 4]

You can use the rotate method for this. This method rotates the position of each element by n. So your examples can be done like this
array.rotate(3).each {|e| e }
and
array.reverse.rotate(1).each {|e| e}
Note: for the second method the parameter to rotate can be derived by finding the negative index of n. So for this the element at index 3 is at index -2 in a length 5 array.

You can do this with upto and downto Fixnum's methods:
array = [0, 1, 2, 3, 4]
last_index = array.size - 1
3.upto last_index do |i|
puts array[i]
end
# => 3, 4
last_index.downto 3 do |i|
puts array[i]
end
# => 4, 3
PS. as speed benchmark, iteration with rotation faster
array.rotate(3).each {|e| puts e}
benchmark:
require 'benchmark'
array = Array.new(10000000) { rand(1...9) }
last_index = array.size - 1
Benchmark.bm do |x|
x.report 'upto' do
10000.upto last_index do |index| a = array[index] + 1; end
end
x.report 'downto' do
last_index.downto 10000 do |index| a = array[index] + 1; end
end
x.report 'rotate' do
array.rotate(10000).each {|e| a = e + 1 }
end
end
# RESULTS:
# user system total real
# upto 0.680000 0.000000 0.680000 ( 0.681932)
# downto 0.680000 0.000000 0.680000 ( 0.679752)
# rotate 0.590000 0.040000 0.630000 ( 0.622901)
but, as memory benchmark, iteration by array indexes less memory hungry, especially on big array sizes:
require 'memory_profiler'
array = Array.new(10000000) { rand(1...9) }
last_index = array.size - 1
{
upto: -> { 10000.upto last_index do |index| a = array[index] + 1; end },
downto: -> { last_index.downto 10000 do |index| a = array[index] + 1; end },
rotate: -> { array.rotate(10000).each {|e| a = e + 1 } },
reverse_rotate: -> { array.reverse.rotate(10000).each {|e| a = e + 1 } }
}.each { |desc, code| puts "#{desc.to_s} => #{MemoryProfiler.report(&code).total_allocated_memsize.to_s}" }
# RESULTS (in bytes):
# upto => 0 # no additional memory allocation
# downto => 0 # no additional memory allocation
# rotate => 80000040 # implicitly copied array 1 time
# reverse_rotate => 160000080 # implicitly copied array 2 times

Related

How can I pass in a block to my "bubble sort" method?

The below code is my newbie take on a bubble sort method.
#For each element in the list, look at that element and the element
#directly to it's right. Swap these two elements so they are in
#ascending order.
def bubble_sort (array)
a = 0
b = 1
until (array.each_cons(2).all? { |a, b| (a <=> b) <= 0}) == true do
sort = lambda {array[a] <=> array[b]}
sort_call = sort.call
loop do
case sort_call
when -1 #don't swap
a += 1
b += 1
break
when 0 #don't swap
a += 1
b += 1
break
when 1 #swap
array.insert(a,array.delete_at(b))
a += 1
b += 1
break
else #end of array, return to start
a = 0
b = 1
break
end
end
end
puts array.inspect
end
array = [4, 2, 5, 6, 3, 23, 5546, 234, 234, 6]
bubble_sort(array)
I want to be able to alter this method so that it takes a block of code as an argument and uses this to determine how it sorts.
For example:
array = ["hello", "my", "name", "is", "daniel"]
bubble_sort(array) {array[#a].length <=> array[#b].length}
(When I've tried this I've turned a and b into instance variables throughout the code.)
I have tried using yield but I get undefined method 'length' for nil:NilClass once the end of the array is reached. I've tried adding in things such as
if array[#b+1] == nil
#a = 0
#b = 1
end
This helps but I still end up with weird problems like infinite loops or not being able to sort more than certain amount of elements.
Long story short, I have been at this for hours. Is there a simple way to do what I want to do? Thanks.
The way you're calling your lambda is a bit odd. It's actually completely unnecessary. I refactored your code and cleaned up a bit of the redundancy. The following works for me:
def sorted?(arr)
arr.each_cons(2).all? { |a, b| (a <=> b) <= 0 }
end
def bubble_sort (arr)
a = 0
b = 1
until sorted?(arr) do
# The yield call here passes `arr[a]` and `arr[b]` to the block.
comparison = if block_given?
yield(arr[a], arr[b])
else
arr[a] <=> arr[b]
end
if [-1, 0, 1].include? comparison
arr.insert(a, arr.delete_at(b)) if comparison == 1
a += 1
b += 1
else
a = 0
b = 1
end
end
arr
end
sample_array = [4, 2, 5, 6, 3, 23, 5546, 234, 234, 6]
# Sanity check:
100.times do
# `a` is the value of `arr[a]` in our function above. Likewise for `b` and `arr[b]`.
print bubble_sort(sample_array.shuffle) { |a, b| a <=> b }, "\n"
end
EDIT
A cleaner version:
# In place swap will be more efficient as it doesn't need to modify the size of the arra
def swap(arr, idx)
raise IndexError.new("Index #{idx} is out of bounds") if idx >= arr.length || idx < 0
temp = arr[idx]
arr[idx] = arr[idx + 1]
arr[idx + 1] = temp
end
def bubble_sort(arr)
loop do
sorted_elements = 0
arr.each_cons(2).each_with_index do |pair, idx|
comparison = if block_given?
yield pair.first, pair.last
else
pair.first <=> pair.last
end
if comparison > 0
swap(arr, idx)
else
sorted_elements += 1
end
end
return arr if sorted_elements >= arr.length - 1
end
end
# A simple test
sample_array = [4, 2, 2, 2, 2, 2, 5, 5, 6, 3, 23, 5546, 234, 234, 6]
sample_str_array = ["a", "ccc", "ccccc"]
100.times do
print bubble_sort(sample_array.shuffle) { |a, b| a <=> b }, "\n"
print bubble_sort(sample_str_array.shuffle) { |a, b| a.length <=> b.length }, "\n"
end
You're not too far off. Just a few things:
Make your function take a block argument
def bubble_sort (array, &block)
Check to see if the user has provided a block
if block_given?
# Call user's comparator block
else
# Use the default behavior
end
Call the user's comparator block
block.call(a, b)
In the user-provided block, accept block params for the elements to compare
bubble_sort(array) {|a,b| a.length <=> b.length}
That should put you in the right ballpark.

Overlap ranges in array

I would like to write a program in ruby 1.9.3 ver. which collects unique value ranges and then calculates amount of numbers in these ranges.
For example lets use 3 ranges (1..3), (6..8) and (2..4). It will return array with two ranges (1..4), (6..8) and amount of numbers - 7.
I wrote the following code:
z= []
def value_ranges(start, finish, z)
range = (start..finish)
arr = z
point = nil
if arr.empty?
point = nil
else
arr.each { |uniq|
if overlap?(uniq,range) == true
point = arr.index(uniq)
break
else
point = nil
end
}
end
if point != nil
if arr[point].first >= start && arr[point].end <= finish
range = (start..finish)
elsif arr[point].first >= start
range = (start..arr[point].end)
elsif arr[point].end <= finish
range = (arr[point].first..finish)
else
range = (arr[point].first..arr[point].end)
end
arr[point] = range
else
arr << range
end
print arr
end
def overlap?(x,y)
(x.first - y.end) * (y.first - x.end) >= 0
end
Problem comes when program meets a range which overlaps two already collected ranges.
For example (1..5) (7..9) (11..19) and the next given range is (8..11).
It should link both ranges and return the following result - (1..5),(7..19).
I don't have an idea how to recheck whole array without creating infinite loop. Also what is the best way to calculate amount of numbers in ranges?
Here are two Ruby-like ways of doing it.
arr = [1..3, 6..8, 2..4]
#1 Efficient approach
First calculate the amalgamated ranges:
a = arr[1..-1].sort_by(&:first).each_with_object([arr.first]) do |r,ar|
if r.first <= ar.last.last
ar[-1] = ar.last.first..[ar.last.last,r.last].max
else
ar << r
end
end
#=> [1..4, 6..8]
Then compute the total number of elements in those ranges:
a.reduce(0) { |tot,r| tot + r.size }
#=> [1..4, 6..8].reduce(0) { |tot,r| tot + r.size }
#=> 7
Explanation
b = arr[1..-1]
#=> [6..8, 2..4]
c = b.sort_by(&:first)
#=> [2..4, 6..8]
enum = c.each_with_object([1..3])
#=> #<Enumerator: [2..4, 6..8]:each_with_object([1..3])>
The contents of the enumerator enum will be passed into the block and assigned to the block variables by Enumerator#each, which will call Array#each. We can see the contents of the enumerator by converting it to an array:
enum.to_a
#=> [[2..4, [1..3]], [6..8, [1..3]]]
and we can use Enumerator#next to step through the enumerator. The first element of the enumerator passed to the block by each is [2..4, [1..3]]. This is assigned to the block variables as follows:
r, ar = enum.next
#=> [2..4, [1..3]]
r #=> 2..4
ar #=> [1..3]
We now perform the block calculation
if r.first <= ar.last.last
#=> 2 <= (1..3).last
#=> 2 <= 3
#=> true
ar[-1] = ar.last.first..[ar.last.last,r.last].max
#=> ar[-1] = 1..[3,4].max
#=> ar[-1] = 1..4
#=> 1..4
else # not executed this time
ar << r
end
This is not so mysterious. So I don't have to keep saying "the last range of ar", let me define:
ar_last = ar.last
#=> 1..3
First of all, because we began by sorting the ranges by the beginning of each range, we know that when each element of enum is passed into the block:
ar_last.first <= r.first
For each element of enum passed into the block for which:
r.first <= ar_last.last
we compare r.last with ar_last.last. There are two possibilities to consider:
r.last <= ar_last.last, in which case the two ranges overlap and therefore ar_last would not change; and
r.last > ar_last.last, in which case the upper end of ar_last must be increased to r.last.
Here,
2 = r.first <= ar_last.last = 3
4 = r.last > ar_last.last = 3
so ar_last is changed from 1..3 to 1..4.
each now passes the last element of enum into the block:
r, ar = enum.next
#=> [6..8, [1..4]]
r #=> 6..8
ar #=> [1..4]
if r.first <= ar.last.last
#=> (6 <= 4) => false this time
...
else # executed this time
ar << r
#=> ar << (6..8)
#=> [1..4, 6..8]
end
and
a = ar #=> [1..4, 6..8]
This time, r.first > ar_last.last, meaning the range r does not overlap ar_last, so we append r to ar, and ar_last now equals r.
Lastly:
a.reduce(0) { |tot,r| tot + r.size }
#=> [1..4, 6..8].reduce(0) { |tot,r| tot + r.size }
#=> 7
which we could alternatively write:
a.map(&:size).reduce(:+)
#2 Easy but inefficient
Here is an easy, but not especially efficient, method that uses Enumerable#slice_when, newly-minted in v2.2.
arr = [(1..3), (6..8), (2..4)]
To calculate the amagamated ranges:
a = arr.flat_map(&:to_a)
.uniq
.sort
.slice_when { |i,j| i+1 != j }
.map { |ar| (ar.first..ar.last) }
#=> [1..4, 6..8]
The total number of elements in those ranges is calculated as in #1
Explanation
Here are the steps:
b = arr.flat_map(&:to_a)
#=> [1, 2, 3, 6, 7, 8, 2, 3, 4]
c = b.uniq
#=> [1, 2, 3, 6, 7, 8, 4]
d = c.sort
#=> [1, 2, 3, 4, 6, 7, 8]
e = d.slice_when { |i,j| i+1 != j }
#=> #<Enumerator: #<Enumerator::Generator:0x007f81629584f0>:each>
a = e.map { |ar| (ar.first..ar.last) }
#=> [1..4, 6..8]
We can see the contents of the enumerator e by converting it to an array:
e.to_a
#=> [[1, 2, 3, 4], [6, 7, 8]]

How can I remove the first 3 duplicate values and return an array with the remaining values?

Given these arrays, how do I remove three occurrences of a value while keeping the fourth or fifth in the array?
[1,5,1,1,1] # => [1,5]
[3,3,3,2,3] # => [3,2]
[3,4,5,3,3] # => [4,5]
[1,1,1,1,1] # => [1,1]
[1,2,2,4,5] # => [1,2,2,4,5]
Here's what I've tried:
array = [1,5,1,1,1]
top3 = array.select { |x| array.count(x) >= 3 }[0..2]
last2 = array - top3
This strategy (and similar) only seem to work when there are three duplicates but not four or five. Are there elegant solutions to this problem?
UPDATE: Thank you for your amazing answers. As a beginning rubyist I learned a lot just from analyzing each response. My question came from a Ruby Koan challenge for a dice program. Here's my complete solution implemented with Abdo's suggestion. I'm sure there are more efficient ways to implement the program :)
def score(dice)
a,b,c,d,e = dice
array = [a,b,c,d,e]
total = 0
triples = array.select {|x| array.count(x) >= 3}[0..2]
singles = array.group_by{|i| i}.values.map{ |a|
a.length > 2 ? a[0, a.length - 3] : a
}.inject([], :+)
# Calculate values for triples
# 1 * 3 = 1000pts
# 2 * 3 = 200pts
# 3 * 3 = 300pts
# 4 * 3 = 400pts
# 5 * 3 = 500pts
# 6 * 3 = 600pts
case triples[0]
when 1 then total += triples[0]*1000
when (2..6) then total += triples[0]*100
end
# Calculate values for singles:
# 1s = 100pts each
# 5s = 50pts each
singles.include? (1) ? singles.select {|x| x == 1 }.each {|x| total += x*100 } : total
singles.include? (5) ? singles.select {|x| x == 5 }.each {|x| total += x*10 } : total
return total
end
puts score([5,1,1, 5, 6]) # 300 points
puts score([]) # 0 points
puts score([1,1,1,5,1]) # 1150 points
puts score([2,3,4,6,2]) # 0 points
puts score([3,4,5,3,3]) # 350 points
puts score([1,5,1,2,4]) # 250 points
array = [1,5,1,1,1]
occurrence = {}
array.select do|a|
if(array.count(a) > 3)
occurrence[a] ||= []
occurrence[a] << a
occurrence[a].count > 3
else
true
end
end
PS: This solution preserves the order of the elements in the original array
Here's a faster solution when the size of the array is large:
(I avoid using count because it would loop through the array in an inner loop)
arr.inject({}) {
|h, i| h[i] ||= 0; h[i] += 1; h
}.collect_concat {|k,v| [k] * (v > 2 ? v - 3 : v) }
Here's the fruity comparison to the other working solutions:
arr = 1000.times.collect { rand(100) }.shuffle
require 'fruity'
compare do
vimsha {
occurrence = {};
arr.select do|a|
if(arr.count(a) > 3)
occurrence[a] ||= []
occurrence[a] << a
occurrence[a].count > 3
else
true
end
end
}
caryswoveland {
arr.uniq.reduce([]) {|a,e| a + [e]*((cnt=arr.count(e)) > 2 ? cnt-3 : cnt)}
}
aruprakshit {
num_to_del = arr.find { |e| arr.count(e) >= 3 }
if !num_to_del.nil?
3.times do
ind = arr.index { |e| e == num_to_del }
arr.delete_at(ind)
end
end
arr
}
# edited as suggested by #CarySwoveland
abdo {
arr.each_with_object(Hash.new {|h,k| h[k]=[]}) {|i,h| h[i] += 1
}.collect_concat { |k,v| [k] * (v > 2 ? v - 3 : v) }
}
broisatse {
arr.group_by{|i| i}.values.map{ |a|
a.length > 2 ? a[0, a.length - 3] : a
}.inject([], :+)
}
end
Here's the comparison result:
Running each test 64 times. Test will take about 48 seconds.
broisatse is faster than abdo by 30.000000000000004% ± 10.0%
abdo is faster than aruprakshit by 4x ± 1.0 (results differ: ...)
aruprakshit is similar to caryswoveland (results differ: ...)
caryswoveland is similar to vimsha (results differ: ...)
Note: I took #aruprakshit's code outside the method so we don't waste time in the method call itself.
When the array's size is increased further:
arr = 1000.times.collect { rand(1000) }.shuffle
we get:
abdo is faster than broisatse by 3x ± 1.0
broisatse is faster than aruprakshit by 6x ± 10.0
aruprakshit is faster than caryswoveland by 2x ± 1.0
caryswoveland is similar to vimsha
Another way, assuming order need not be preserved (which is consistent with a comment by the asker):
array = [1,2,4,1,2,1,2,1,1,4]
array.uniq.reduce([]) {|a,e| a + [e]*((cnt=array.count(e)) > 2 ? cnt-3 : cnt)}
#=> [1, 1, 4, 4]
Try something like:
a.group_by{|i| i}.values.map{|a| a[0, a.length % 3]}.inject([], :+)
This will remove all triplets from the array. If you want to remove only the first triplet, then do:
a.group_by{|i| i}.values.map{|a| a.length > 2 ? a[0, a.length - 3] : a }.inject([], :+)
Note: This might mess up the order of the array:
[1,2,1,2,3] #=> [1,1,2,2,3]
Let me know if you need to keep the order and, if so, which elements need to be removed if there are more than three, e.g. what should say: [1,1,2,1,1,] - [1,2] or [2,1]?
x.group_by{|i| i }.values.select{|a| a.size >= 3 }.each{|a| c=[3,a.size].min; x.delete_if{|e| a[0]==e && (c-=1)>=0 } }
It will remove the first [3,a.size].min occurrences of a[0] from the input x where a is, for example, [1,1,1,1] for x = [1,2,1,1,1]
I'd do as below :
def del_first_three(a)
num_to_del = a.find { |e| a.count(e) >= 3 }
return a if num_to_del.nil?
3.times do
ind = a.index { |e| e == num_to_del }
a.delete_at(ind)
end
a
end
del_first_three([3,4,5,3,3]) # => [4, 5]
del_first_three([1,5,1,1,1]) # => [5, 1]
del_first_three([1,2,2,4,5]) # => [1, 2, 2, 4, 5]

Calculating Median in Ruby

How do I calculate the median of an array of numbers using Ruby?
I am a beginner and am struggling with handling the cases of the array being of odd and even length.
Here is a solution that works on both even and odd length array and won't alter the array:
def median(array)
return nil if array.empty?
sorted = array.sort
len = sorted.length
(sorted[(len - 1) / 2] + sorted[len / 2]) / 2.0
end
Similar to nbarraille's, but I find it a bit easier to keep track of why this one works:
class Array
def median
sorted = self.sort
half_len = (sorted.length / 2.0).ceil
(sorted[half_len-1] + sorted[-half_len]) / 2.0
end
end
half_len = number of elements up to and including (for array with odd number of items) middle of array.
Even simpler:
class Array
def median
sorted = self.sort
mid = (sorted.length - 1) / 2.0
(sorted[mid.floor] + sorted[mid.ceil]) / 2.0
end
end
If by calculating Median you mean this
Then
a = [12,3,4,5,123,4,5,6,66]
a.sort!
elements = a.count
center = elements/2
elements.even? ? (a[center] + a[center+1])/2 : a[center]
def median(array) #Define your method accepting an array as an argument.
array = array.sort #sort the array from least to greatest
if array.length.odd? #is the length of the array odd?
array[(array.length - 1) / 2] #find value at this index
else array.length.even? #is the length of the array even?
(array[array.length/2] + array[array.length/2 - 1])/2.to_f
#average the values found at these two indexes and convert to float
end
end
More correct solution with handling edge cases:
class Array
def median
sorted = self.sort
size = sorted.size
center = size / 2
if size == 0
nil
elsif size.even?
(sorted[center - 1] + sorted[center]) / 2.0
else
sorted[center]
end
end
end
There is a specs to prove:
describe Array do
describe '#median' do
subject { arr.median }
context 'on empty array' do
let(:arr) { [] }
it { is_expected.to eq nil }
end
context 'on 1-element array' do
let(:arr) { [5] }
it { is_expected.to eq 5 }
end
context 'on 2-elements array' do
let(:arr) { [1, 2] }
it { is_expected.to eq 1.5 }
end
context 'on odd-size array' do
let(:arr) { [100, 5, 2, 12, 1] }
it { is_expected.to eq 5 }
end
context 'on even-size array' do
let(:arr) { [7, 100, 5, 2, 12, 1] }
it { is_expected.to eq 6 }
end
end
end
I like to use Refinements, which is a safe way to Monkey Patch the ruby classes without collateral effects over the system.
The usage become much more cleaner than a new method.
With the Refinements you can monkey patch the Array class, implement the Array#median and this method will only be available inside the scope of the class that is using the refinement! :)
Refinements
module ArrayRefinements
refine Array do
def median
return nil if empty?
sorted = sort
(sorted[(length - 1) / 2] + sorted[length / 2]) / 2.0
end
end
end
class MyClass
using ArrayRefinements
# You can use the Array#median as you wish here
def test(array)
array.median
end
end
MyClass.new.test([1, 2, 2, 2, 3])
=> 2.0
def median(array)
half = array.sort!.length / 2
array.length.odd? ? array[half] : (array[half] + array[half - 1]) / 2
end
*If the length is even, you must add the middle point plus the middle point - 1 to account for the index starting at 0
def median(arr)
sorted = arr.sort
if sorted == []
return nil
end
if sorted.length % 2 != 0
result = sorted.length / 2 # 7/2 = 3.5 (rounded to 3)
return sorted[result] # 6
end
if sorted.length % 2 == 0
result = (sorted.length / 2) - 1
return (sorted[result] + sorted[result+1]) / 2.0 # (4 + 5) / 2
end
end
p median([5, 0, 2, 6, 11, 10, 9])
Here's a solution:
app_arry = [2, 3, 4, 2, 5, 6, 16].sort
# check array isn't empty
if app_arry.empty? || app_arry == ""
puts "Sorry, This will not work."
return nil
end
length = app_arry.length
puts "Array length = #{length}"
puts "Array = #{app_arry}"
if length % 2 == 0
# even number of elements
puts "median is #{(app_arry[length/2].to_f + app_arry[(length-1)/2].to_f)/2}"
else
# odd number of elements
puts "median is #{app_arry[(length-1)/2]}"
end
OUTPUT
Array length = 7
Array = [2, 3, 4, 2, 5, 6, 16]
median is 2
def median(array, already_sorted=false)
return nil if array.empty?
array = array.sort unless already_sorted
m_pos = array.size / 2
return array.size % 2 == 1 ? array[m_pos] : mean(array[m_pos-1..m_pos])
end
There are many ways to do this, but for both performance and reliability, I suggest using the enumerable-statistics library created by Ruby committer mrkn.
https://github.com/mrkn/enumerable-statistics
require 'enumerable/statistics'
ary = [1,2,3,3,4]
ary.mean # => 2.6
ary.median # => 3
I think it's good:
#!/usr/bin/env ruby
#in-the-middle value when odd or
#first of second half when even.
def median(ary)
middle = ary.size/2
sorted = ary.sort_by{ |a| a }
sorted[middle]
end
or
#in-the-middle value when odd or
#average of 2 middle when even.
def median(ary)
middle = ary.size/2
sorted = ary.sort_by{ |a| a }
ary.size.odd? ? sorted[middle] : (sorted[middle]+sorted[middle-1])/2.0
end
I used sort_by rather than sort because it's faster: Sorting an array in descending order in Ruby.

How can I merge and order multiple lists together using Ruby?

I have 2 lists that have dates and data. Each list is in the proper order as noted by the sequence number. Now I need to merge the 2 lists together and keep everything in the correct order.
For example:
List A
20101001 A data 1 seq1
20101001 A data 2 seq2
20101005 A data 3 seq3
List B
20101001 B data 1 seq1
20101003 B data 2 seq2
etc...
I need the new list to look like this:
20101001 A data 1 seq1
20101001 A data 2 seq2
20101001 B data 1 seq3
20101003 B data 2 seq4
20101005 A data 3 seq5
2 things that I thought of is merging the lists together and applying the sequence number prior to inserting them into a db or I can insert them into the db with the current sequence and pull them back out again to merge them together, but that seems like an extra step and kludgy.
Any ideas on the best way to go about this?
Assuming your lists are in Ruby Arrays, and the objects in the lists have attributes defined (such as obj.sequence_number), one way to merge and sort the lists would be:
First merge the lists as a union:
#merged_list = #list_a | #list_b
Then sort the merged_list with the appropriate sorting rule:
#merged_list.sort! {|a, b| a.date <=> b.date # or whatever your sorting rule is... }
Edit:
Once the merged array is sorted, you can re-define the sequence_number:
#merged_list.each_with_index {|obj, index| obj.sequence_number = "seq#{index+1}"}
Edit:
Same thing applies if your objects in the lists are themselves just simple arrays:
#merged_list.sort! {|a, b| a[0] <=> b[0] # or whatever your sorting rule is... }
#merged_list.each_with_index {|obj, index| obj[2] = "seq#{index+1}"}
Try this:
(listA + listB).sort!{|a, b| a.sequence_no <=> b.sequence_no}
This is an algorithm for merging an arbitrary number of sorted lists in more or less linear time:
def merge_sorted(*lists)
# the lists will be modified, so make (shallow) copies
lists = lists.map(&:dup)
result = []
loop do
# ignore lists that have been exhausted
lists = lists.reject(&:empty?)
# we're done if all lists have been exhausted
break if lists.empty?
# find the list with the smallest first element
top = lists.inject do |candidate, other|
candidate.first < other.first ? candidate : other
end
result << top.shift
end
result
end
list1 = [1, 2, 5, 6, 9]
list2 = [2, 3, 4, 11, 13]
list3 = [1, 2, 2, 2, 3]
p merge_sorted(list1, list2, list3)
# => [1, 1, 2, 2, 2, 2, 2, 3, 3, 4, 5, 6, 9, 11, 13]
For each iteration it finds the list with the smallest first element, and shifts this element off of it onto the results list. It does this until all lists are empty.
I say more or less linear time since it's actually O(n × m) where n is the number of lists and m is the total number of elements in the lists but I think this can safely be simplified to O(m) for most cases since n will be small in comparison to m.
This uses with_index which is a nice way to add an index value to an iterator:
result = (list_a + list_b).sort_by { |a| a[0 .. -2] }.map.with_index { |a, i| a[0 .. -2] + (1 + i).to_s }
puts result
# >> 20101001 A data 1 seq1
# >> 20101001 A data 2 seq2
# >> 20101001 B data 1 seq3
# >> 20101003 B data 2 seq4
# >> 20101005 A data 3 seq5
Here's some variations with benchmarks:
require 'benchmark'
list_a = [
'20101001 A data 1 seq1',
'20101001 A data 2 seq2',
'20101005 A data 3 seq3'
]
list_b = [
'20101001 B data 1 seq1',
'20101003 B data 2 seq2'
]
# #1
result = (list_a + list_b).sort_by { |a| a[0 .. -2] }.map.with_index { |a, i| a[0 .. -2] + (1 + i).to_s }
result # => ["20101001 A data 1 seq1", "20101001 A data 2 seq2", "20101001 B data 1 seq3", "20101003 B data 2 seq4", "20101005 A data 3 seq5"]
# #2
result = (list_a + list_b).map{ |r| r[0 .. -2] }.sort.map.with_index { |a, i| a + (1 + i).to_s }
result # => ["20101001 A data 1 seq1", "20101001 A data 2 seq2", "20101001 B data 1 seq3", "20101003 B data 2 seq4", "20101005 A data 3 seq5"]
# #3
i = 0
result = (list_a + list_b).map{ |r| r[0 .. -2] }.sort.map { |a| i += 1; a + i.to_s }
result # => ["20101001 A data 1 seq1", "20101001 A data 2 seq2", "20101001 B data 1 seq3", "20101003 B data 2 seq4", "20101005 A data 3 seq5"]
# #4
i = 0; result = (list_a + list_b).sort.map { |a| i += 1; a[-1] = i.to_s; a }
result # => ["20101001 A data 1 seq1", "20101001 A data 2 seq2", "20101001 B data 1 seq3", "20101003 B data 2 seq4", "20101005 A data 3 seq5"]
n = 75000
Benchmark.bm(7) do |x|
x.report('#1') { n.times { (list_a + list_b).sort_by { |a| a[0 .. -2] }.map.with_index { |a, i| a[0 .. -2] + (1 + i).to_s } } }
x.report('#2') { n.times { (list_a + list_b).map{ |r| r[0 .. -2] }.sort.map.with_index { |a, i| a + (1 + i).to_s } } }
x.report('#3') { n.times { i = 0; (list_a + list_b).map{ |r| r[0 .. -2] }.sort.map { |a| i += 1; a + i.to_s } } }
x.report('#4') { n.times { i = 0; (list_a + list_b).sort.map { |a| i += 1; a[-1] = i.to_s } } }
end
# >> user system total real
# >> #1 1.150000 0.000000 1.150000 ( 1.147090)
# >> #2 0.880000 0.000000 0.880000 ( 0.880038)
# >> #3 0.720000 0.000000 0.720000 ( 0.727135)
# >> #4 0.580000 0.000000 0.580000 ( 0.572688)
It's good to benchmark.

Resources