bamfcsv 0.3.1 → 0.3.2

data/README

@@ -23,5 +23,13 @@ irb(1.9.2): table.first["foo"]
 irb(1.9.2): table[1]["bar"]
 ==========> "4"
 
+DOUBLE-AND/ALSO it supports custom field separators!!
+
+irb(1.9.2): BAMFCSV.parse(<<EOF, separator: '|')
+         ?> foo|bar
+         ?> baz|quux
+         ?> EOF
+==========> [["foo", "bar"], ["baz", "quux"]]
+
 
 (*) Results may vary.

data/ext/bamfcsv/bamfcsv_ext.c

@@ -28,10 +28,11 @@ bool quotes_end_line(char* cur) {
   return *(cur-1) == '"' || (*(cur-1) == '\r' && *(cur-2) == '"');
 }
 
-VALUE bamfcsv_parse_string(VALUE self, VALUE string) {
+VALUE bamfcsv_parse_string(VALUE self, VALUE string, VALUE rstr_sep) {
   char *buf = RSTRING_PTR(string);
   long bufsize = RSTRING_LEN(string);
   rb_encoding *enc = rb_enc_from_index(ENCODING_GET(string));
+  char separator = *RSTRING_PTR(rstr_sep);
 
   unsigned long num_rows = 1, cell_count = 1;
   int quote_count = 0, quotes_matched = 1;
@@ -62,7 +63,7 @@ VALUE bamfcsv_parse_string(VALUE self, VALUE string) {
 
     if (quotes_matched) { 
 
-      if (*cur == ',') {
+      if (*cur == separator) {
         
         if (quote_count && *(cur-1) != '"')
           rb_raise(BAMFCSV_MalformedCSVError_class, "Unclosed quoted field on line %lu, cell %lu.", num_rows, cell_count);
@@ -124,7 +125,7 @@ void Init_bamfcsv() {
 
   BAMFCSV_module = rb_define_module("BAMFCSV");
   VALUE bamfcsv_singleton_class = rb_singleton_class(BAMFCSV_module);
-  rb_define_private_method(bamfcsv_singleton_class, "__parse_string", bamfcsv_parse_string, 1);
+  rb_define_private_method(bamfcsv_singleton_class, "__parse_string", bamfcsv_parse_string, 2);
 
   BAMFCSV_MalformedCSVError_class = rb_define_class_under(BAMFCSV_module, "MalformedCSVError", rb_eRuntimeError);
 }

data/lib/bamfcsv.rb

@@ -14,7 +14,10 @@ module BAMFCSV
     # copy the pointer, not the contents. So we make a copy, parse
     # that, and throw away the copy.
     copy = "" + csv_str
-    matrix = __parse_string(copy)
+    separator = opts.fetch(:separator, ',')
+    raise InvalidSeparator, "Separator must be exactly one character long" if separator.size != 1
+    raise InvalidSeparator, "Separator cannot be '\"'" if separator == '"'
+    matrix = __parse_string(copy, separator)
     if opts[:headers]
       Table.new(matrix)
     else
@@ -22,4 +25,5 @@ module BAMFCSV
     end
   end
 
+  class InvalidSeparator < StandardError; end
 end

data/lib/bamfcsv/version.rb

@@ -1,3 +1,3 @@
 module BAMFCSV
-  VERSION = "0.3.1"
+  VERSION = "0.3.2"
 end

data/spec/fixtures/pipe-delimited.csv

@@ -0,0 +1,2 @@
+foo|bar
+pipe|delimited

data/spec/lib/bamfcsv_spec.rb

@@ -38,6 +38,28 @@ describe BAMFCSV do
       BAMFCSV.read("spec/fixtures/terminated-with-cr.csv").should == [["a"],["b"]]
     end
 
+    describe "Alternate separators" do
+      it "raises BAMFCSV::InvalidSeparator if :separator is not exactly one character long" do
+        expect { BAMFCSV.read("spec/fixtures/pipe-delimited.csv", :separator => '') }.should raise_error(BAMFCSV::InvalidSeparator)
+        expect { BAMFCSV.read("spec/fixtures/pipe-delimited.csv", :separator => 'as') }.should raise_error(BAMFCSV::InvalidSeparator)
+      end
+
+      it "raises BAMFCSV::InvalidSeparator if :separator is a double-quote" do
+        expect { BAMFCSV.read("spec/fixtures/pipe-delimited.csv", :separator => '"') }.should raise_error(BAMFCSV::InvalidSeparator)
+      end
+
+      it "accepts a single character, non-quote :separator option" do
+        parsed = BAMFCSV.read("spec/fixtures/pipe-delimited.csv", :separator => '|')
+        parsed.should == [["foo", "bar"], ["pipe", "delimited"]]
+      end
+
+      it "works with tables too" do
+        table = BAMFCSV.read("spec/fixtures/pipe-delimited.csv", :headers => true, :separator => '|')
+        table.first["foo"].should == "pipe"
+        table.first["bar"].should == "delimited"
+      end
+    end
+
     it "raises Errno::ENOENT when the file does not exist" do
       expect do
         BAMFCSV.read("spec/fixtures/this-file-does-not-not-exist.csv")
@@ -250,5 +272,29 @@ CSV
         BAMFCSV.parse(csv, :headers => true).first.inspect.should == inspected
       end
     end
+    describe "Alternate separators" do
+      it "raises BAMFCSV::InvalidSeparator if :separator is not exactly one character long" do
+        expect { BAMFCSV.parse("1,2", :separator => '') }.should raise_error(BAMFCSV::InvalidSeparator)
+        expect { BAMFCSV.parse("1,2", :separator => 'as') }.should raise_error(BAMFCSV::InvalidSeparator)
+      end
+
+      it "raises BAMFCSV::InvalidSeparator if :separator is a double-quote" do
+        expect { BAMFCSV.parse("1,2", :separator => '"') }.should raise_error(BAMFCSV::InvalidSeparator)
+      end
+
+      it "accepts a single character, non-quote :separator option" do
+        semicolon = BAMFCSV.parse("foo;bar\n1;2", :separator => ';')
+        semicolon.should == [["foo", "bar"], ["1", "2"]]
+        pipe = BAMFCSV.parse("foo|bar\n1|2", :separator => '|')
+        pipe.should == [["foo", "bar"], ["1", "2"]]
+      end
+
+      it "works with tables too" do
+        table = BAMFCSV.parse("foo;bar\nbaz;qux", :headers => true, :separator => ';')
+        table.first["foo"].should == "baz"
+        table.first["bar"].should == "qux"
+      end
+    end
   end
+
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bamfcsv
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.2
   prerelease: 
 platform: ruby
 authors:
@@ -10,11 +10,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-08-10 00:00:00.000000000Z
+date: 2011-08-27 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
-  requirement: &2157181100 !ruby/object:Gem::Requirement
+  requirement: &2156352680 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -22,10 +22,10 @@ dependencies:
         version: 2.5.0
   type: :development
   prerelease: false
-  version_requirements: *2157181100
+  version_requirements: *2156352680
 - !ruby/object:Gem::Dependency
   name: fuubar
-  requirement: &2157180600 !ruby/object:Gem::Requirement
+  requirement: &2156352180 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -33,10 +33,10 @@ dependencies:
         version: 0.0.2
   type: :development
   prerelease: false
-  version_requirements: *2157180600
+  version_requirements: *2156352180
 - !ruby/object:Gem::Dependency
   name: rake-compiler
-  requirement: &2157180140 !ruby/object:Gem::Requirement
+  requirement: &2156351720 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -44,7 +44,7 @@ dependencies:
         version: 0.7.1
   type: :development
   prerelease: false
-  version_requirements: *2157180140
+  version_requirements: *2156351720
 description: BAMFCSV parses csv like a BAMF. BAMF!!
 email:
 - jon@thinkrelevance.com
@@ -71,6 +71,7 @@ files:
 - spec/fixtures/empty.csv
 - spec/fixtures/escapes.csv
 - spec/fixtures/one-column.csv
+- spec/fixtures/pipe-delimited.csv
 - spec/fixtures/terminated-with-cr.csv
 - spec/fixtures/test.csv
 - spec/lib/bamfcsv_spec.rb
@@ -108,6 +109,7 @@ test_files:
 - spec/fixtures/empty.csv
 - spec/fixtures/escapes.csv
 - spec/fixtures/one-column.csv
+- spec/fixtures/pipe-delimited.csv
 - spec/fixtures/terminated-with-cr.csv
 - spec/fixtures/test.csv
 - spec/lib/bamfcsv_spec.rb