csv-utils 0.3.7 → 0.3.12

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '0248ce0ef1db03a751c864fdf52194b25a21f21ef90cbeaf8d0ae07075cbaf51'
-  data.tar.gz: c48e4f907963b1ace89fecb095f2a7ee17ec1670db2371dd4e015005800b757b
+  metadata.gz: d7b43801f0f7e62d34d8a77508964f0b6ea054a3ac2c8d17bf42d22aa0b245bd
+  data.tar.gz: ccd3282e4c7b1e1985439e24a159fe09c570b18446adee2b8db4776bc6024852
 SHA512:
-  metadata.gz: 848fa57de005d00bdf24cd07b533680cd9307c07a19b2fb955391737e36407dfc6be22edef55742fb0c90e8ed5273d4c584840e7d7a657404c1242947a88692b
-  data.tar.gz: 71209d88db0d67ed606e4db56e3cf667081074523373951fea44698537af0661cf63c79eb9b971419bb13bfc2a8b1b9c10c2f89e911b913b4f6c4bc8d00ee061
+  metadata.gz: 1684319de3c4f728f55e1c201790dd5cc3aa110e09d14a4043fcdf3b70c02568fe898c4f2d7b839e633f461294f9870fae0a8cf2d3d785d173aa66f9c8ba2ab1
+  data.tar.gz: 6e7689f6f544e5280dbf144dc5a23a4b088aeff6c06931c0a856f6cd16353048e4ed343da0406550d2e7f48abd8af6e1ef31f6114ec9f3435aa56815c5fea24b

data/.ruby-version

@@ -1 +1 @@
-2.6.3
+3.0.0

data/Gemfile.lock

@@ -1,50 +1,52 @@
 GEM
   remote: http://rubygems.org/
   specs:
-    ast (2.4.1)
-    diff-lcs (1.3)
-    docile (1.3.2)
+    ast (2.4.2)
+    diff-lcs (1.4.4)
+    docile (1.3.5)
     inheritance-helper (0.1.5)
-    parallel (1.19.2)
-    parser (2.7.1.4)
+    parallel (1.20.1)
+    parser (3.0.0.0)
       ast (~> 2.4.1)
     rainbow (3.0.0)
-    rake (13.0.1)
-    regexp_parser (1.7.1)
+    rake (13.0.3)
+    regexp_parser (2.1.1)
     rexml (3.2.4)
-    rspec (3.9.0)
-      rspec-core (~> 3.9.0)
-      rspec-expectations (~> 3.9.0)
-      rspec-mocks (~> 3.9.0)
-    rspec-core (3.9.2)
-      rspec-support (~> 3.9.3)
-    rspec-expectations (3.9.2)
+    rspec (3.10.0)
+      rspec-core (~> 3.10.0)
+      rspec-expectations (~> 3.10.0)
+      rspec-mocks (~> 3.10.0)
+    rspec-core (3.10.1)
+      rspec-support (~> 3.10.0)
+    rspec-expectations (3.10.1)
       diff-lcs (>= 1.2.0, < 2.0)
-      rspec-support (~> 3.9.0)
-    rspec-mocks (3.9.1)
+      rspec-support (~> 3.10.0)
+    rspec-mocks (3.10.2)
       diff-lcs (>= 1.2.0, < 2.0)
-      rspec-support (~> 3.9.0)
-    rspec-support (3.9.3)
-    rubocop (0.86.0)
+      rspec-support (~> 3.10.0)
+    rspec-support (3.10.2)
+    rubocop (1.11.0)
       parallel (~> 1.10)
-      parser (>= 2.7.0.1)
+      parser (>= 3.0.0.0)
       rainbow (>= 2.2.2, < 4.0)
-      regexp_parser (>= 1.7)
+      regexp_parser (>= 1.8, < 3.0)
       rexml
-      rubocop-ast (>= 0.0.3, < 1.0)
+      rubocop-ast (>= 1.2.0, < 2.0)
       ruby-progressbar (~> 1.7)
-      unicode-display_width (>= 1.4.0, < 2.0)
-    rubocop-ast (0.0.3)
-      parser (>= 2.7.0.1)
-    ruby-progressbar (1.10.1)
-    simplecov (0.18.5)
+      unicode-display_width (>= 1.4.0, < 3.0)
+    rubocop-ast (1.4.1)
+      parser (>= 2.7.1.5)
+    ruby-progressbar (1.11.0)
+    simplecov (0.21.2)
       docile (~> 1.1)
       simplecov-html (~> 0.11)
-    simplecov-html (0.12.2)
-    unicode-display_width (1.7.0)
+      simplecov_json_formatter (~> 0.1)
+    simplecov-html (0.12.3)
+    simplecov_json_formatter (0.1.2)
+    unicode-display_width (2.0.0)
 
 PLATFORMS
-  ruby
+  x86_64-darwin-20
 
 DEPENDENCIES
   inheritance-helper
@@ -54,4 +56,4 @@ DEPENDENCIES
   simplecov
 
 BUNDLED WITH
-   1.17.3
+   2.2.3

data/bin/csv-validator

@@ -35,20 +35,24 @@ id_column_name = ARGV[1]
 headers = csv.shift
 strip_bom!(headers[0])
 
-id_column_num = nil
-if id_column_name
-  unless headers.include?(id_column_name)
-    $stderr.puts("header #{id_column_name} not found in current set of headers")
-    exit 1
-  end
-
-  id_column_num = headers.index(id_column_name)
+id_column_name ||= headers[0]
+unless headers.include?(id_column_name)
+  $stderr.puts("header #{id_column_name} not found in current set of headers")
+  exit 1
 end
 
+id_column_num = headers.index(id_column_name)
+
 out = nil
-if id_column_num
-  out = CSV.open('utf8-correctsion.csv', 'wb')
-  out << [id_column_name, 'Row', 'Col', 'Header', 'Value']
+out_proc = Proc.new do |row|
+  out ||=
+    begin
+      out = CSV.open('utf8-correctsion.csv', 'wb')
+      out << [id_column_name, 'Row', 'Col', 'Header', 'Value']
+      out
+    end
+
+  out << row
 end
 
 csv_lineno = 1
@@ -66,7 +70,7 @@ while (row = csv.shift)
     $stderr.puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: none UTF-8 characters found in \"#{col}\""
     if (col_utf8_encoded = convert_to_utf8(col, detect_encoding(col)))
       puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: converted to UTF-8 from #{detect_encoding(col)} \"#{col_utf8_encoded}\""
-      out << [row[id_column_num], csv_lineno, (idx + 1), headers[idx], col_utf8_encoded]
+      out_proc.call [row[id_column_num], csv_lineno, (idx + 1), headers[idx], col_utf8_encoded]
     else
       $stderr.puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: unknown character encoding"
     end

data/csv-utils.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |s|
   s.name        = 'csv-utils'
-  s.version     = '0.3.7'
+  s.version     = '0.3.12'
   s.licenses    = ['MIT']
   s.summary     = 'CSV Utils'
   s.description = 'Tools for debugging malformed CSV files'

data/lib/csv-utils.rb

@@ -2,6 +2,7 @@ require 'csv'
 
 # Collection of tools for working with CSV files.
 module CSVUtils
+  autoload :CSVCompare, 'csv_utils/csv_compare'
   autoload :CSVExtender, 'csv_utils/csv_extender'
   autoload :CSVIterator, 'csv_utils/csv_iterator'
   autoload :CSVOptions, 'csv_utils/csv_options'

data/lib/csv_utils/csv_compare.rb

@@ -0,0 +1,83 @@
+# frozen_string_literal: true
+
+# CSVUtils::CSVCompare purpose is to determine which rows in the secondary_data_file need to be created, deleted or updated
+# **requires both CSV files to be sorted on the same columns, CSVUtils::CSVSort can accomplish this
+# In order to receive updates, update_comparison_columns must configured or use inheritance and change the update_row? method
+class CSVUtils::CSVCompare
+  # primary_data_file is the source of truth
+  # compare_proc used to compare the id column(s)
+  # update_comparison_columns column(s) to compare for equality, ex: updated_at, timestamp, hash
+  #  caveat: update_comparison_columns need to be in both csv files
+  attr_reader :primary_data_file,
+              :update_comparison_columns,
+              :compare_proc
+
+  def initialize(primary_data_file, update_comparison_columns=nil, &block)
+    @primary_data_file = primary_data_file
+    @update_comparison_columns = update_comparison_columns
+    @compare_proc = block
+  end
+
+  def compare(secondary_data_file)
+    src = CSV.open(primary_data_file)
+    src_headers = src.shift
+    dest = CSV.open(secondary_data_file)
+    dest_headers = dest.shift
+
+    read_next_src = true
+    read_next_dest = true
+
+    while(!src.eof? || !dest.eof?)
+      src_record = next_record_from_file(src_headers, src) if read_next_src
+      dest_record = next_record_from_file(dest_headers, dest) if read_next_dest
+
+      if ! src_record
+        read_next_src = false
+        read_next_dest = true
+
+        yield :delete, dest_record
+      elsif ! dest_record
+        read_next_src = true
+        read_next_dest = false
+
+        yield :create, src_record
+      elsif compare_proc.call(src_record, dest_record) == 0
+        read_next_src = true
+        read_next_dest = true
+
+        yield(:update, src_record) if update_row?(src_record, dest_record)
+      elsif compare_proc.call(src_record, dest_record) > 0
+        read_next_src = false
+        read_next_dest = true
+
+        yield :delete, dest_record
+      else
+        read_next_src = true
+        read_next_dest = false
+
+        yield :create, src_record
+      end
+    end
+
+    src.close
+    dest.close
+  end
+
+  private
+
+  def next_record_from_file(headers, file)
+    return nil if file.eof?
+
+    Hash[headers.zip(file.shift)]
+  end
+
+  def update_row?(src_record, dest_record)
+    return false unless update_comparison_columns
+
+    update_comparison_columns.each do |column_name|
+      return true unless src_record[column_name] == dest_record[column_name]
+    end
+
+    false
+  end
+end

data/lib/csv_utils/csv_iterator.rb

@@ -14,7 +14,7 @@ class CSVUtils::CSVIterator
     end
 
     def to_pretty_s
-      reject { |_, v| v.strip.empty? }
+      reject { |_, v| v.nil? || v.strip.empty? }
         .each_with_index
         .map { |(k, v), idx| sprintf('  %-3d %s: %s', idx+1, k, v) }
         .join("\n") + "\n"
@@ -43,6 +43,26 @@ class CSVUtils::CSVIterator
     end
   end
 
+  def headers
+    first.keys
+  end
+
+  def to_hash(key, value = nil)
+    raise("header #{key} not found in #{headers}") unless headers.include?(key)
+    raise("headers #{value} not found in #{headers}") if value && !headers.include?(value)
+
+    value_proc =
+      if value
+        proc { |row| row[value] }
+      else
+        proc { |row| yield(row) }
+      end
+
+    each_with_object({}) do |row, hsh|
+      hsh[row[key]] = value_proc.call(row)
+    end
+  end
+
   private
 
   def strip_bom!(col)

data/lib/csv_utils/csv_report.rb

@@ -9,17 +9,18 @@ module CSVUtils
         if csv.is_a?(String)
           @must_close = true
           mode = csv_options.delete(:mode) || 'wb'
-          CSV.open(csv, mode, csv_options)
+          CSV.open(csv, mode, **csv_options)
         else
           @must_close = false
           csv
         end
 
-      generate(headers, &block) if block
+      add_headers(headers) if headers
+
+      generate(&block) if block
     end
 
-    def generate(headers = nil)
-      add_headers(headers) if headers
+    def generate
       yield self
       close if @must_close
     end

data/lib/csv_utils/csv_sort.rb

@@ -25,9 +25,9 @@ class CSVUtils::CSVSort
   private
 
   def merge_sort_csv_files(src_csv_file1, src_csv_file2, dest_csv_file)
-    src1 = CSV.open(src_csv_file1, 'rb', csv_options)
-    src2 = CSV.open(src_csv_file2, 'rb', csv_options)
-    dest = CSV.open(dest_csv_file, 'wb', csv_options)
+    src1 = CSV.open(src_csv_file1, 'rb', **csv_options)
+    src2 = CSV.open(src_csv_file2, 'rb', **csv_options)
+    dest = CSV.open(dest_csv_file, 'wb', **csv_options)
 
     if @headers
       dest << @headers
@@ -66,7 +66,7 @@ class CSVUtils::CSVSort
   end
 
   def create_sorted_csv_part_files(batch_size, &block)
-    src = CSV.open(csv_file, 'rb', csv_options)
+    src = CSV.open(csv_file, 'rb', **csv_options)
 
     @headers = src.shift if has_headers
 
@@ -74,7 +74,7 @@ class CSVUtils::CSVSort
     create_batch_part_proc = Proc.new do
       batch.sort!(&block)
       @csv_part_files << "#{new_csv_file}.part.#{@csv_part_files.size}"
-      CSV.open(@csv_part_files.last, 'wb', csv_options) do |csv|
+      CSV.open(@csv_part_files.last, 'wb', **csv_options) do |csv|
         csv << @headers if @headers
         batch.each { |row| csv << row }
       end
@@ -107,6 +107,10 @@ class CSVUtils::CSVSort
       File.unlink(csv_part_file2)
     end
 
-    FileUtils.mv(@csv_part_files.last, new_csv_file)
+    if @csv_part_files.size > 0
+      FileUtils.mv(@csv_part_files.last, new_csv_file)
+    else
+      FileUtils.cp(@csv_file, new_csv_file)
+    end
   end
 end

data/lib/csv_utils/csv_wrapper.rb

@@ -20,7 +20,7 @@ class CSVUtils::CSVWrapper
   def open(csv, mode, csv_options)
     if csv.is_a?(String)
       @close_when_done = true
-      @csv = CSV.open(csv, mode, csv_options)
+      @csv = CSV.open(csv, mode, **csv_options)
     else
       @close_when_done = false
       @csv = csv

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: csv-utils
 version: !ruby/object:Gem::Version
-  version: 0.3.7
+  version: 0.3.12
 platform: ruby
 authors:
 - Doug Youch
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2020-12-05 00:00:00.000000000 Z
+date: 2021-05-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: inheritance-helper
@@ -49,6 +49,7 @@ files:
 - bin/csv-validator
 - csv-utils.gemspec
 - lib/csv-utils.rb
+- lib/csv_utils/csv_compare.rb
 - lib/csv_utils/csv_extender.rb
 - lib/csv_utils/csv_iterator.rb
 - lib/csv_utils/csv_options.rb
@@ -77,7 +78,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.8
+rubygems_version: 3.2.3
 signing_key:
 specification_version: 4
 summary: CSV Utils