Removing duplicates from a csv file using ruby -

i have csv file following data

sno scenario    result  description 1   sce_1   pass    pass 2   sce_2   pass    pass 1   sce_1   fail    failed 

in scenario have 2 same serial numbers. see row has result pass , remove rest of duplicate rows.

i have tried following still not able it!

csv.open('new.csv', 'w') |csv|    csv.read('merged_files.csv').uniq!{|x| x[1]}.each |row|        csv << row    end end 

can me in getting logic!

for purposes of illustration, i've added fourth row table:

require 'csv'  arr = csv.read("x.csv")   #=> [["sno", "scenario", "result", "description"],   #    ["1", "sce_1", "pass", "pass"],   #    ["2", "sce_2", "pass", "pass"],   #    ["1", "sec_1", "fail", "pass"],   #    ["3", "sec_3", "fail", "pass"]] 

you can remove unwanted elements follows:

arr[1..-1].group_by(&:first).map { |_,a|   (a.size > 1) ? a.reject { |e| e[2]=="fail" } : }   #=> [[["1", "sce_1", "pass", "pass"]],   #    [["2", "sce_2", "pass", "pass"]],   #    [["3", "sec_3", "fail", "pass"]]] 

the steps:

h = arr[1..-1].group_by(&:first)   #=> {"1"=>[["1", "sce_1", "pass", "pass"],   #          ["1", "sec_1", "fail", "pass"]],   #    "2"=>[["2", "sce_2", "pass", "pass"]],   #    "3"=>[["3", "sec_3", "fail", "pass"]]}  h.map { |_,a| (a.size > 1) ? a.reject { |e| e[2]=="fail" } : }   #=> [[["1", "sce_1", "pass", "pass"]],   #    [["2", "sce_2", "pass", "pass"]],   #    [["3", "sec_3", "fail", "pass"]]] 

if, given sno/scenario there @ 1 "pass" row, can use enumerable#flat_map instead:

a = h.flat_map { |_,a| (a.size > 1) ? a.reject { |e| e[2]=="fail" } : }   #=> [["1", "sce_1", "pass", "pass"],   #    ["2", "sce_2", "pass", "pass"],   #    ["3", "sec_3", "fail", "pass"]] 

if wish add header row:

a.unshift(arr.first)   #=> [["sno", "scenario", "result", "description"],   #    ["1", "sce_1", "pass", "pass"],   #    ["2", "sce_2", "pass", "pass"],   #    ["3", "sec_3", "fail", "pass"]] 

if want exclude "fail" rows, if there no correponding "pass" row (as sno == "3"), can this:

h.flat_map { |_,a| a.reject { |e| e[2]=="fail" } }   #=> [["1", "sce_1", "pass", "pass"],   #    ["2", "sce_2", "pass", "pass"]]