Searching through two multidimensional arrays and grouping together similar subarrays - ruby

Question

I am trying to search through two multidimensional arrays to find any elements in common in a given subarray and then put the results in a third array where the entire subarrays with similar elements are grouped together (not just the similar elements).
The data is imported from two CSVs:
require 'csv'
array = CSV.read('primary_csv.csv')
#=> [["account_num", "account_name", "primary_phone", "second_phone", "status],
#=> ["11111", "John Smith", "8675309", " ", "active"],
#=> ["11112", "Tina F.", "5551234", "5555678" , "disconnected"],
#=> ["11113", "Troy P.", "9874321", " ", "active"]]
# and so on...
second_array = CSV.read('customer_service.csv')
#=> [["date", "name", "agent", "call_length", "phone", "second_phone", "complaint"],
#=> ["3/1/15", "Mary ?", "Bob X", "5:00", "5551234", " ", "rude"],
#=> ["3/2/15", "Mrs. Smith", "Stew", "1:45", "9995678", "8675309" , "says shes not a customer"]]
# and so on...
If any number is present as an element in a subarray on both primary.csv and customer_service.csv, I want that entire subarray (as opposed to just the common elements), put into a third array, results_array. The desire output based upon the above sample is:
results_array = [["11111", "John Smith", "8675309", " ", "active"],
["3/2/15", "Mrs. Smith", "Stew", "1:45", "9995678", "8675309" , "says shes not a customer"]] # and so on...
I then want to export the array into a new CSV, where each subarray is its own row of the CSV. I intend to iterate over each subarray by joining it with a , to make it comma delimited and then put the results into a new CSV:
results_array.each do {|j| j.join(",")}
File.open("results.csv", "w") {|f| f.puts results_array}
#=> 11111,John Smith,8675309, ,active
#=> 3/2/15,Mrs. Smith,Stew,1:45,9995678,8675309,says shes not a customer
# and so on...
How can I achieve the desired output? I am aware that the final product will look messy because similar data (for example, phone number) will be in different columns. But I need to find a way to generally group the data together.

Answer 1

Suppose a1 and a2 are the two arrays (excluding header rows).
Code
def combine(a1, a2)
h2 = a2.each_with_index
.with_object(Hash.new { |h,k| h[k] = [] }) { |(arr,i),h|
arr.each { |e| es = e.strip; h[es] << i if number?(es) } }
a1.each_with_object([]) do |arr, b|
d = arr.each_with_object([]) do |str, d|
s = str.strip
d.concat(a2.values_at(*h2[s])) if number?(s) && h2.key?(s)
end
b << d.uniq.unshift(arr) if d.any?
end
end
def number?(str)
str =~ /^\d+$/
end
Example
Here is your example, modified somewhat:
a1 = [
["11111", "John Smith", "8675309", "", "active" ],
["11112", "Tina F.", "5551234", "5555678", "disconnected"],
["11113", "Troy P.", "9874321", "", "active" ]
]
a2 = [
["3/1/15", "Mary ?", "Bob X", "5:00", "5551234", "", "rude"],
["3/2/15", "Mrs. Smith", "Stew", "1:45", "9995678", "8675309", "surly"],
["3/7/15", "Cher", "Sonny", "7:45", "9874321", "8675309", "Hey Jude"]
]
combine(a1, a2)
#=> [[["11111", "John Smith", "8675309", "",
# "active"],
# ["3/2/15", "Mrs. Smith", "Stew", "1:45",
# "9995678", "8675309", "surly"],
# ["3/7/15", "Cher", "Sonny", "7:45",
# "9874321", "8675309", "Hey Jude"]
# ],
# [["11112", "Tina F.", "5551234", "5555678",
# "disconnected"],
# ["3/1/15", "Mary ?", "Bob X", "5:00",
# "5551234", "", "rude"]
# ],
# [["11113", "Troy P.", "9874321", "",
# "active"],
# ["3/7/15", "Cher", "Sonny", "7:45",
# "9874321", "8675309", "Hey Jude"]
# ]
# ]
Explanation
First, we define a helper:
def number?(str)
str =~ /^\d+$/
end
For example:
number?("8675309") #=> 0 ("truthy)
number?("3/1/15") #=> nil
Now index a2 on the values that represent numbers:
h2 = a2.each_with_index
.with_object(Hash.new { |h,k| h[k] = [] }) { |(arr,i),h|
arr.each { |e| es = e.strip; h[es] << i if number?(es) } }
#=> {"5551234"=>[0], "9995678"=>[1], "8675309"=>[1, 2], "9874321"=>[2]}
This says, for example, that the "numeric" field "8675309" is contained in elements at offsets 1 and 2 of a2 (i.e, for Mrs. Smith and Cher).
We can now simply run through the elements of a1 looking for matches.
The code:
arr.each_with_object([]) do |str, d|
s = str.strip
d.concat(a2.values_at(*h2[s])) if number?(s) && h2.key?(s)
end
steps through the elements of arr, assigning each to the block variable str. For example, if arr holds the first element of a1 str will in turn equals "11111", "John Smith", and so on. After s = str.strip, this says that if a s has a numerical representation and there is a matching key in h2, the (initially empty) array d is concatenated with the elements of a2 given by the value of h2[s].
After completing this loop we see if d contains any elements of a2:
b << d.uniq.unshift(arr) if d.any?
If it does, we remove duplicates, prepend the array with arr and save it to b.
Note that this allows one element of a2 to match multiple elements of a1.