How to combine arrays of similar data in Ruby?

Question

I am working on a program that will compare two .csv files. After extracting the relevant data from one of the csv files into an array of arrays, I need to combine related entries. For example, I would want to turn this array:

[["11/13/15", ["4001", "1392"], "INBOUND"], 
["11/13/15", ["4090", "540"], "INBOUND"], 
["11/13/15", ["1139", "162"], "INBOUND"], 
["11/13/15", ["1158", "64"], "INBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["4055", "448"], "OUTBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["1139", "162"], "OUTBOUND"], 
["11/13/15", ["1158", "64"], "OUTBOUND"], 
["11/13/15", ["4091", "520"], "OUTBOUND"]]

into this:

[["11/13/15", ["4001", "1392"], "INBOUND"], 
["11/13/15", ["4090", "540"], "INBOUND"], 
["11/13/15", ["1139", "162"], "INBOUND"], 
["11/13/15", ["1158", "64"], "INBOUND"], 
["11/13/15", ["4055", "1152"], "OUTBOUND"], 
["11/13/15", ["1139", "162"], "OUTBOUND"], 
["11/13/15", ["1158", "64"], "OUTBOUND"], 
["11/13/15", ["4091", "520"], "OUTBOUND"]]

For some element of the array, if its items at [0], [1][0], and [2] match those of another one, then create a new item (array) with its item at [1][1] being the sum of all the items at [1][1] and delete the old arrays. If it would be easier, I can change the way the relevant data is extracted so that the item at [1] is not an array and each row has 4 items instead of 3.

The data will be sorted so that it looks like the top array if printed, so yes (if I understand your question). — Sean Lerner
– Sean Lerner, Commented Nov 15, 2011 at 21:55
Assuming tokland's answer was realy what you wanted, your question has a typo for the value at [4][1][1] of the resulting array, which is the sole crucial value. It should be 1152, not 1115. I must say your question is sloppy. — sawa
– sawa, Commented Nov 15, 2011 at 22:24
@sawa Didn't notice that typo. Thanks for helping clarify my sentence as well. — Sean Lerner
– Sean Lerner, Commented Nov 17, 2011 at 20:54

tokland · Accepted Answer · 2011-11-15 22:01:05Z

2

I assume that the elements to group are consecutive so we can use Enumerable#chunk. Functional approach:

grouped_xs = xs.chunk { |date, (id1, id2), direction| [date, id1, direction] }
grouped_xs.map do |(date, id1, direction), ary|
  id2_sum = ary.map { |date, (id1, id2), direction| id2.to_i }.inject(:+)
  [date, id1, id2_sum.to_s, direction]
end

Output (you wanted 4 elements in the output array, right?):

[["11/13/15", "4001", "1392", "INBOUND"],
 ["11/13/15", "4090", "540", "INBOUND"],
 ["11/13/15", "1139", "162", "INBOUND"],
 ["11/13/15", "1158", "64", "INBOUND"],
 ["11/13/15", "4055", "1152", "OUTBOUND"],
 ["11/13/15", "1139", "162", "OUTBOUND"],
 ["11/13/15", "1158", "64", "OUTBOUND"],
 ["11/13/15", "4091", "520", "OUTBOUND"]]

edited Nov 15, 2011 at 22:01

answered Nov 15, 2011 at 21:55

tokland

68.2k13 gold badges151 silver badges174 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

tokland Over a year ago

@Sean: you are welcome. Just as a general advice, I think it's better not to rush into selecting an answer too quickly, somebody may come up with a better solution :-)

Sean Vikoren Over a year ago

That is some very compact Ruby @tokland :)

tokland Over a year ago

@SeanVikoren: Funcional solutions tend to be very compact. Sometimes they are a bit harder to read than imperative ones, but conceptually they are very nice (like putting together lego pieces instead of building everything from scratch).

Sean Lerner Over a year ago

@tokland I will keep that in mind for future questions (which I am sure I will have many).

Sean Vikoren Over a year ago

@tokland: It does have a mesmerizing beauty.

Oleksandr Skrypnyk · Accepted Answer · 2011-11-16 08:27:18Z

2

And just for example - my one-liner (works with both 1.8 and 1.9 rubies):

table = [["11/13/15", ["4001", "1392"], "INBOUND"], 
["11/13/15", ["4090", "540"], "INBOUND"], 
["11/13/15", ["1139", "162"], "INBOUND"], 
["11/13/15", ["1158", "64"], "INBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["4055", "448"], "OUTBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["1139", "162"], "OUTBOUND"], 
["11/13/15", ["1158", "64"], "OUTBOUND"], 
["11/13/15", ["4091", "520"], "OUTBOUND"]]

result = table.group_by {|a, (b, c), d| [a, [b], d]}.map {|k, v| k[1] << v.map {|a| a[1][1].to_i}.inject(:+).to_s; k}

edited Nov 16, 2011 at 8:27

answered Nov 15, 2011 at 23:11

Oleksandr Skrypnyk

2,8301 gold badge23 silver badges26 bronze badges

1 Comment

pguardiario Over a year ago

I especially like "group_by {|a, (b, c), d|" - very nice.

Sean Vikoren · Accepted Answer · 2011-11-15 22:07:23Z

0

This should do it:

def lookup(list, id, direction)
  index = nil
  list.each_with_index do |e, i|
    if (id == e[1][0]) and (e[2] == direction)
      index = i
      break
    end
  end
  index
end

b = []

a.each do |e|
  id = e[1][0]
  direction = e[2]
  i = lookup(b, id, direction)
  if i.nil?
    b << e
  else
    count = e[1][1].to_i
    sum = count + b[i][1][1].to_i
    b[i][1][1] = sum.to_s
  end
end

b.each{|e| p e}

Output:

["11/13/15", ["4001", "1392"], "INBOUND"]
["11/13/15", ["4090", "540"], "INBOUND"]
["11/13/15", ["1139", "162"], "INBOUND"]
["11/13/15", ["1158", "64"], "INBOUND"]
["11/13/15", ["4055", "1152"], "OUTBOUND"]
["11/13/15", ["1139", "162"], "OUTBOUND"]
["11/13/15", ["1158", "64"], "OUTBOUND"]
["11/13/15", ["4091", "520"], "OUTBOUND"]

edited Nov 15, 2011 at 22:07

answered Nov 15, 2011 at 21:57

Sean Vikoren

1,0941 gold badge13 silver badges24 bronze badges

Comments

sawa · Accepted Answer · 2011-11-15 22:30:23Z

h = Hash.new(0)
[["11/13/15", ["4001", "1392"], "INBOUND"], 
["11/13/15", ["4090", "540"], "INBOUND"], 
["11/13/15", ["1139", "162"], "INBOUND"], 
["11/13/15", ["1158", "64"], "INBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["4055", "448"], "OUTBOUND"], 
["11/13/15", ["4055", "352"], "OUTBOUND"], 
["11/13/15", ["1139", "162"], "OUTBOUND"], 
["11/13/15", ["1158", "64"], "OUTBOUND"], 
["11/13/15", ["4091", "520"], "OUTBOUND"]]
.each{|a, (b, c), d| h[[a, b, d]] += c.to_i}
p h.map{|(a, b, d), c| [a, [b, c], d]}

will give:

[["11/13/15", ["4001", 1392], "INBOUND"],
["11/13/15", ["4090", 540], "INBOUND"],
["11/13/15", ["1139", 162], "INBOUND"],
["11/13/15", ["1158", 64], "INBOUND"],
["11/13/15", ["4055", 1152], "OUTBOUND"],
["11/13/15", ["1139", 162], "OUTBOUND"],
["11/13/15", ["1158", 64], "OUTBOUND"],
["11/13/15", ["4091", 520], "OUTBOUND"]]

Collectives™ on Stack Overflow

How to combine arrays of similar data in Ruby?

4 Answers 4

5 Comments

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

5 Comments

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related