diff --git a/lib/csv.rb b/lib/csv.rb
index 6456da8..b22c06e 100644
--- a/lib/csv.rb
+++ b/lib/csv.rb
@@ -1019,6 +1019,7 @@ class MalformedCSVError < RuntimeError; end
# :skip_blanks:: +false+
# :force_quotes:: +false+
# :skip_lines:: +nil+
+ # :liberal_parsing:: +false+
#
DEFAULT_OPTIONS = {
col_sep: ",",
@@ -1033,6 +1034,7 @@ class MalformedCSVError < RuntimeError; end
skip_blanks: false,
force_quotes: false,
skip_lines: nil,
+ liberal_parsing: false,
}.freeze
#
@@ -1499,6 +1501,10 @@ def self.table(path, options = Hash.new)
# a comment. If the passed object does
# not respond to match,
# ArgumentError is thrown.
+ # :liberal_parsing:: When set to a +true+ value, CSV will
+ # attempt to parse input not conformant
+ # with RFC 4180, such as double quotes
+ # in unquoted fields.
#
# See CSV::DEFAULT_OPTIONS for the default settings.
#
@@ -1622,6 +1628,8 @@ def header_converters
def skip_blanks?() @skip_blanks end
# Returns +true+ if all output fields are quoted. See CSV::new for details.
def force_quotes?() @force_quotes end
+ # Returns +true+ if illegal input is handled. See CSV::new for details.
+ def liberal_parsing?() @liberal_parsing end
#
# The Encoding CSV is parsing or writing in. This will be the Encoding you
@@ -1860,12 +1868,12 @@ def shift
end
elsif part[0] == @quote_char
# If we are starting a new quoted column
- if part[-1] != @quote_char || part.count(@quote_char) % 2 != 0
+ if part.count(@quote_char) % 2 != 0
# start an extended column
csv << part[1..-1]
csv.last << @col_sep
in_extended_col = true
- else
+ elsif part[-1] == @quote_char
# regular quoted column
csv << part[1..-2]
if csv.last =~ @parsers[:stray_quote]
@@ -1873,6 +1881,11 @@ def shift
"Missing or stray quote in line #{lineno + 1}"
end
csv.last.gsub!(@quote_char * 2, @quote_char)
+ elsif @liberal_parsing
+ csv << part
+ else
+ raise MalformedCSVError,
+ "Missing or stray quote in line #{lineno + 1}"
end
elsif part =~ @parsers[:quote_or_nl]
# Unquoted field with bad characters.
@@ -1880,7 +1893,11 @@ def shift
raise MalformedCSVError, "Unquoted fields do not allow " +
"\\r or \\n (line #{lineno + 1})."
else
- raise MalformedCSVError, "Illegal quoting in line #{lineno + 1}."
+ if @liberal_parsing
+ csv << part
+ else
+ raise MalformedCSVError, "Illegal quoting in line #{lineno + 1}."
+ end
end
else
# Regular ole unquoted field.
@@ -1945,7 +1962,7 @@ def inspect
str << " encoding:" << @encoding.name
# show other attributes
%w[ lineno col_sep row_sep
- quote_char skip_blanks ].each do |attr_name|
+ quote_char skip_blanks liberal_parsing ].each do |attr_name|
if a = instance_variable_get("@#{attr_name}")
str << " " << attr_name << ":" << a.inspect
end
@@ -2079,6 +2096,7 @@ def init_parsers(options)
# store the parser behaviors
@skip_blanks = options.delete(:skip_blanks)
@field_size_limit = options.delete(:field_size_limit)
+ @liberal_parsing = options.delete(:liberal_parsing)
# prebuild Regexps for faster parsing
esc_row_sep = escape_re(@row_sep)
diff --git a/test/csv/test_features.rb b/test/csv/test_features.rb
index 37fdab8..a558875 100755
--- a/test/csv/test_features.rb
+++ b/test/csv/test_features.rb
@@ -142,6 +142,29 @@ def test_skip_blanks
assert_equal(3, count)
end
+ def test_liberal_parsing
+ input = '"Johnson, Dwayne",Dwayne "The Rock" Johnson'
+ assert_raise(CSV::MalformedCSVError) do
+ CSV.parse_line(input)
+ end
+ assert_equal(["Johnson, Dwayne", 'Dwayne "The Rock" Johnson'],
+ CSV.parse_line(input, liberal_parsing: true))
+
+ input = '"quoted" field'
+ assert_raise(CSV::MalformedCSVError) do
+ CSV.parse_line(input)
+ end
+ assert_equal(['"quoted" field'],
+ CSV.parse_line(input, liberal_parsing: true))
+
+ assert_raise(CSV::MalformedCSVError) do
+ CSV.parse_line('is,this "three," or four,fields', liberal_parsing: true)
+ end
+
+ assert_equal(["is", 'this "three', ' or four"', "fields"],
+ CSV.parse_line('is,this "three, or four",fields', liberal_parsing: true))
+ end
+
def test_csv_behavior_readers
%w[ unconverted_fields return_headers write_headers
skip_blanks force_quotes ].each do |behavior|