Class: ETL::Processor::EscapeCsvProcessor

Inherits:
Processor show all
Defined in:
lib/etl/processor/escape_csv_processor.rb

Instance Attribute Summary (collapse)

Instance Method Summary (collapse)

Constructor Details

- (EscapeCsvProcessor) initialize(control, configuration)

Initialize the processor.

Configuration options:

  • :source_file: The file to load data from

  • :target_file: The file to write data to

  • :file: short-cut which will set the same value to both source_file and target_file

Raises:



23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 23

def initialize(control, configuration)
  super
  if configuration[:file]
    @use_temp_file = true
    configuration[:source_file] = configuration[:file]
    configuration[:target_file] = configuration[:file] + '.tmp'
  end
  path = Pathname.new(configuration[:source_file])
  @source_file = path.absolute? ? path : Pathname.new(File.dirname(File.expand_path(configuration[:source_file]))) + path
  path = Pathname.new(configuration[:target_file])
  @target_file = path.absolute? ? path : Pathname.new(File.dirname(File.expand_path(configuration[:target_file]))) + path
  @filters = configuration[:filters] || [{:replace => '\"', :result => '""'}]
  @charcount = configuration[:charcount]
  raise ControlError, "Source file must be specified" if @source_file.nil?
  raise ControlError, "Target file must be specified" if @target_file.nil?
  raise ControlError, "Source and target file cannot currently point to the same file" if @source_file == @target_file
end

Instance Attribute Details

- (Object) charcount (readonly)

Returns the value of attribute charcount



15
16
17
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 15

def charcount
  @charcount
end

- (Object) filters (readonly)

Returns the value of attribute filters



14
15
16
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 14

def filters
  @filters
end

- (Object) source_file (readonly)

The file to load from



8
9
10
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 8

def source_file
  @source_file
end

- (Object) target_file (readonly)

The file to write to



10
11
12
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 10

def target_file
  @target_file
end

- (Object) use_temp_file (readonly)

whether to use a temporary file or not



12
13
14
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 12

def use_temp_file
  @use_temp_file
end

Instance Method Details

- (Object) process

Execute the processor



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
 
# File 'lib/etl/processor/escape_csv_processor.rb', line 42

def process
  reader = File.open(@source_file, 'r')
  writer = File.open(@target_file, 'w')

  reader.each_line do |line|
    reading = line
    @filters.each do |filter|
      if (!filter[:replace].nil? &&
          !filter[:result].nil?)
        result = reading.gsub(Regexp.new(filter[:replace]), filter[:result])
        reading = result
      end
    end unless @filters.nil?
    @charcount.each do |count|
      if (!count[:char].nil? &&
          !count[:count].nil?)
        c = reading.count count[:char]
        if c != count[:count]
          reading = nil
        end
      end
    end unless @charcount.nil?
    writer.write(reading) unless reading.nil?
  end

  reader.close
  writer.close

  if use_temp_file
    FileUtils.rm(source_file)
    FileUtils.mv(target_file,source_file)
  end
end