Identify and resolve bottleneck with csv parsing

1 files changed, 28 insertions, 0 deletions

diff --git a/bin/csv-benchmark b/bin/csv-benchmark
new file mode 100755
index 0000000..f7573f2
--- /dev/null
+++ b/bin/csv-benchmark
@@ -0,0 +1,28 @@
+#!/usr/bin/env ruby
+
+require 'bundler/inline'
+
+gemfile do
+  source 'https://rubygems.org'
+
+  gem 'benchmark-ips', '~> 2.8'
+  gem 'fastcsv', '~> 0.0'
+  gem 'fastest-csv'
+end
+
+require 'benchmark/ips'
+require 'csv'
+require 'fastcsv'
+require 'fastest-csv'
+
+csv = "\"spandx\",\"0.0.0\",\"MIT\""
+
+Benchmark.ips do |x|
+  x.report("CSV.parse") { CSV.parse(csv)[0] }
+  x.report("csv.split") { csv.split(',', 3) }
+  x.report("csv.split-with-slice") { csv.chomp.split(',', 3).slice(1...-1) }
+  x.report("csv.regex") { csv.scan(/"(\S+)","*(\d+.\d+.\d+)","(\S+)"/)[0] }
+  x.report("csv.fastcsv") { FastCSV.raw_parse(csv) { |x| x } }
+  x.report("csv.fastestcsv") { FastestCSV.parse_line(csv) }
+  x.compare!
+end