csv-utils 0.3.3 → 0.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a28b89b25aa7a6a90137b799c580aa9060d783f481000fd173281bc7e2367baf
-  data.tar.gz: 7317ece2b8970a816b5e4b5b84623a2c966c3a2f7afa7c2dd30de831f7e3ac64
+  metadata.gz: 8e649e8e220856a0676e01ef58811ab9df3822a1757eae4e776a45256063c510
+  data.tar.gz: fad8bcac595659bc5d91b4ebc19ee8f18ebbdf910568e75ff812768e1c23515f
 SHA512:
-  metadata.gz: e057cc9795559c630e4b939c67d00547a25e0cf83ed75b15713a6c381677a718dd2ac06f975486e5cdf45a1cbfd962caf2aa13105e57a544aa4b6fd177926445
-  data.tar.gz: ee7552522db9fb683b8cbd73a94706a0a41e92135aed0c2ebec6c70754f1bcad210234b8e5b18da87322bbc6737b6f323c515a1920f09fda3230a5cdef2393d1
+  metadata.gz: 4e4a26c3494c466a0099f72db534b8bd73910444c6eadf1b1faa78213d225b6752e9e8b60818bbeabe57f4542bd235da1969a9b6a3bad968328622531f89f077
+  data.tar.gz: d35d7a46f4f58a0fe87e10c6bcfec3ededc840d5d3c7ce0f1e5162e22f6f9ddb1516ed23a59b93877529ae326d597dfbde5089d9676393d488e476f6a9ad0924

data/bin/csv-validator

@@ -25,11 +25,32 @@ def detect_encoding(col)
   CharDet.detect(col)['encoding']
 end
 
+def strip_bom!(col)
+  col.sub!("\xEF\xBB\xBF".force_encoding('ASCII-8BIT'), '')
+end
+
 csv = CSV.open(ARGV[0], 'rb')
-out = CSV.open(ARGV[1], 'wb') if ARGV[1]
+id_column_name = ARGV[1]
 
 headers = csv.shift
-out << headers if out
+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)
+end
+
+out = nil
+if id_column_num
+  out = CSV.open('utf8-correctsion.csv', 'wb')
+  out << [id_column_name, 'Row', 'Col', 'Header', 'Value']
+end
+
 csv_lineno = 1
 
 while (row = csv.shift)
@@ -39,21 +60,17 @@ while (row = csv.shift)
     $stderr.puts "row(#{csv_lineno}): invalid number of columns, expected #{headers.size} got #{row.size}"
   end
 
-  converted = false
   row.each_with_index do |col, idx|
     next if utf8?(col)
 
     $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)))
-      converted = true
       puts "row(#{csv_lineno}),col(#{idx + 1}) #{headers[idx]}: converted to UTF-8 from #{detect_encoding(col)} \"#{col_utf8_encoded}\""
-      row[idx] = col_utf8_encoded
+      out << [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
   end
-
-  out << row if out && converted
 end
 
 csv.close

data/csv-utils.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |s|
   s.name        = 'csv-utils'
-  s.version     = '0.3.3'
+  s.version     = '0.3.4'
   s.licenses    = ['MIT']
   s.summary     = 'CSV Utils'
   s.description = 'Tools for debugging malformed CSV files'

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: csv-utils
 version: !ruby/object:Gem::Version
-  version: 0.3.3
+  version: 0.3.4
 platform: ruby
 authors:
 - Doug Youch