CODEWITHSUNDEEP
rubyarrayssearch
Norto23
Norto23

Reputation: 2269

ruby searching array for keywords

I am parsing a large CSV file in a ruby script and need to find the closest match for a title from some search keys. The search keys maybe one or more values and the values may not exactly match as per below (should be close)

search_keys = ["big", "bear"]

A large array containing data that I need to search through, only want to search on the title column:

array = [
          ["id", "title",            "code", "description"],
          ["1",  "once upon a time", "3241", "a classic story"],
          ["2",  "a big bad wolf",   "4235", "a little scary"],
          ["3",  "three big bears",  "2626", "a heart warmer"]
        ]

In this case I would want it to return the row ["3", "three big bears", "2626", "a heart warmer"] as this is the closest match to my search keys.

I want it to return the closest match from the search keys given.

Is there any helpers/libraries/gems I can use? Anyone done this before??

Upvotes: 5

Views: 1171

Answers (5)

peter
peter

Reputation: 42182

Here is my one-line shot

p array.find_all {|a|a.join.scan(/#{search_keys.join("|")}/).length==search_keys.length}
=>[["3", "three big bears", "2626", "a heart warmer"]]

to get all the rows in order of number of matches

p array.drop(1).sort_by {|a|a.join.scan(/#{search_keys.join("|")}/).length}.reverse

Anyone knows how to combine the last solution so that the rows that contain none of the keys are dropped and to keep it concise as is ?

Upvotes: 1

Soup
Soup

Reputation: 1719

You could probably write it in a more succinct way...

array = [
          ["id", "title",            "code", "description"],
          ["1",  "once upon a time", "3241", "a classic story"],
          ["2",  "a big bad wolf",   "4235", "a little scary"],
          ["3",  "three big bears",  "2626", "a heart warmer"]
        ]
search_keys = ["big", "bear"]


def sift(records, target_field, search_keys)
    # find target_field index
    target_field_index = nil
    records.first.each_with_index do |e, i|
        if e == target_field
            target_field_index = i
            break
        end
    end
    if target_field_index.nil?
        raise "Target field was not found"
    end

    # sums up which records have a match and how many keys they match
    # key => val = record => number of keys matched
    counter = Hash.new(0) # each new hash key is init'd with value of 0

    records.each do |record| # look at all our given records
        search_keys.each do |key| # check each search key on the field
            if record[target_field_index].include?(key)
                counter[record] += 1 # found a key, init to 0 if required and increment count
            end
        end
    end

    # find the result with the most search key matches
    top_result = counter.to_a.reduce do |top, record|
        if record[1] > top[1] # [0] = record, [1] = key hit count
            top = record # set to new top
        end
        top # continue with reduce
    end.first # only care about the record (not the key hit count)
end


puts "Top result: #{sift array, 'title', search_keys}"
# => Top result: ["3", "three big bears", "2626", "a heart warmer"]

Upvotes: 1

TomDunning
TomDunning

Reputation: 4877

This works. Will find and return an array of matched* rows as result.

*matched rows = a row where the id, title, code or description match ANY of the provided seach_keys. incl partial searches such as 'bear' in 'bears'

result = []
array.each do |a|
    a.each do |i|
        search_keys.each do |k|
            result << a if i.include?(k)
        end
    end
end
result.uniq!

Upvotes: 1

Amol Pujari
Amol Pujari

Reputation: 2349

I am worried, this task should be handled to any search engine at db level or similar, no point fetching data in app and do searching across columns/rows etc, should be expensive. but for now here is the plain simple approach :)

array = [
          ["id", "title",            "code", "description"],
          ["1",  "once upon a time", "3241", "a classic story"],
          ["2",  "a big bad wolf",   "4235", "a little scary"],
          ["3",  "three big bears",  "2626", "a heart warmer"]
        ]


h = {}

search_keys = ["big", "bear"]

array[1..-1].each do |rec|
  rec_id = rec[0].to_i

  search_keys.each do |key|
    if rec[1].include? key
      h[rec_id] = h[rec_id] ? (h[rec_id]+1) : 1
    end
  end
end

closest = h.keys.first

h.each do |rec, count| 
  closest = rec if h[closest] < h[rec]
end

array[closest] # => desired output :)

Upvotes: 2

M.ElSaka
M.ElSaka

Reputation: 1294

I think you can do it by your self and no need to use any gems! This may be close to what you need; searching in the array for the keys and set a rank for each found element.

result = []
array.each do |ar|
    rank = 0
    search_keys.each do |key|
        if ar[1].include?(key)
            rank += 1
        end
    end

    if rank > 0
        result << [rank, ar]
    end 
end

This code can be written better than the above, but i wanted to show you the details.

Upvotes: 1

Related Questions