Programming Tutorials

Open and manipulate CSV files in Ruby

By: Vijay in Ruby Tutorials on 2009-03-03  

1. To Open a CSV formatted file for reading or writing use 

open(path, mode, fs = nil, rs = nil, &block)

For reading.

EXAMPLE 1

  CSV.open('csvfile.csv', 'r') do |row|
    p row
  end

EXAMPLE 2

  reader = CSV.open('csvfile.csv', 'r')
  row1 = reader.shift
  row2 = reader.shift
  if row2.empty?
    p 'row2 not find.'
  end
  reader.close

ARGS

  filename: filename to parse.
  col_sep: Column separator.  ?, by default.  If you want to separate
    fields with semicolon, give ?; here.
  row_sep: Row separator.  nil by default.  nil means "\r\n or \n".  If you
    want to separate records with \r, give ?\r here.

RETURNS

  reader instance.  To get parse result, see CSV::Reader#each.

For writing.

EXAMPLE 1

  CSV.open('csvfile.csv', 'w') do |writer|
    writer << ['r1c1', 'r1c2']
    writer << ['r2c1', 'r2c2']
    writer << [nil, nil]
  end

EXAMPLE 2

  writer = CSV.open('csvfile.csv', 'w')
  writer << ['r1c1', 'r1c2'] << ['r2c1', 'r2c2'] << [nil, nil]
  writer.close

ARGS

  filename: filename to generate.
  col_sep: Column separator.  ?, by default.  If you want to separate
    fields with semicolon, give ?; here.
  row_sep: Row separator.  nil by default.  nil means "\r\n or \n".  If you
    want to separate records with \r, give ?\r here.

RETURNS

  writer instance.  See CSV::Writer#<< and CSV::Writer#add_row to know how
  to generate CSV string.

2. Convert a line from cells data to string. Consider using CSV.generate_line instead. To generate multi-row CSV string, see EXAMPLE below.

generate_row(src, cells, out_dev, fs = nil, rs = nil)

EXAMPLE

  row1 = ['a', 'b']
  row2 = ['c', 'd']
  row3 = ['e', 'f']
  src = [row1, row2, row3]
  buf = ''
  src.each do |row|
    parsed_cells = CSV.generate_row(row, 2, buf)
    puts "Created #{ parsed_cells } cells."
  end
  p buf

ARGS

  src: an Array of String to be converted to CSV string.  Must respond to
    'size' and '[](idx)'.  src[idx] must return String.
  cells: num of cells in a line.
  out_dev: buffer for generated CSV string.  Must respond to '<<(string)'.
  col_sep: Column separator.  ?, by default.  If you want to separate
    fields with semicolon, give ?; here.
  row_sep: Row separator.  nil by default.  nil means "\r\n or \n".  If you
    want to separate records with \r, give ?\r here.

RETURNS

  parsed_cells: num of converted cells.

3. parse(str_or_readable, fs = nil, rs = nil) {|row| ...} Parse lines from given string or stream. Return rows as an Array of Arrays.

# File csv.rb, line 115
  def CSV.parse(str_or_readable, fs = nil, rs = nil, &block)
    if File.exist?(str_or_readable)
      STDERR.puts("CSV.parse(filename) is deprecated." +
        "  Use CSV.open(filename, 'r') instead.")
      return open_reader(str_or_readable, 'r', fs, rs, &block)
    end
    if block
      CSV::Reader.parse(str_or_readable, fs, rs) do |row|
        yield(row)
      end
      nil
    else
      CSV::Reader.create(str_or_readable, fs, rs).collect { |row| row }
    end
  end

4. parse_line(src, fs = nil, rs = nil) Parse a line from given string. Bear in mind it parses ONE LINE. Rest of the string is ignored for example "a,b\r\nc,d" => ['a", 'b"] and the second line 'c,d" is ignored.

If you don't know whether a target string to parse is exactly 1 line or not, use CSV.parse_row instead of this method.

# File csv.rb, line 137
  def CSV.parse_line(src, fs = nil, rs = nil)
    fs ||= ','
    if fs.is_a?(Fixnum)
      fs = fs.chr
    end
    if !rs.nil? and rs.is_a?(Fixnum)
      rs = rs.chr
    end
    idx = 0
    res_type = :DT_COLSEP
    row = []
    begin
      while res_type == :DT_COLSEP
        res_type, idx, cell = parse_body(src, idx, fs, rs)
        row << cell
      end
    rescue IllegalFormatError
      return []
    end
    row
  end

5. parse_row(src, idx, out_dev, fs = nil, rs = nil) Parse a line from string. Consider using CSV.parse_line instead. To parse lines in CSV string, see EXAMPLE below.

EXAMPLE

  src = "a,b\r\nc,d\r\ne,f"
  idx = 0
  begin
    parsed = []
    parsed_cells, idx = CSV.parse_row(src, idx, parsed)
    puts "Parsed #{ parsed_cells } cells."
    p parsed
  end while parsed_cells > 0

ARGS

  src: a CSV data to be parsed.  Must respond '[](idx)'.
    src[](idx) must return a char. (Not a string such as 'a', but 97).
    src[](idx_out_of_bounds) must return nil.  A String satisfies this
    requirement.
  idx: index of parsing location of 'src'.  0 origin.
  out_dev: buffer for parsed cells.  Must respond '<<(aString)'.
  col_sep: Column separator.  ?, by default.  If you want to separate
    fields with semicolon, give ?; here.
  row_sep: Row separator.  nil by default.  nil means "\r\n or \n".  If you
    want to separate records with \r, give ?\r here.

RETURNS

  parsed_cells: num of parsed cells.
  idx: index of next parsing location of 'src'.