require 'csv'
module Expertus
class Dataset < Array
attr_accessor :data_type, :udc_id, :filename_suffix
# knows its common_udc_id and data type
def initialize *args
CSV::Converters[:non_octal_integer] = lambda { |f|
begin
e = f.encode(CSV::ConverterEncoding)
unless e == "0"
if e.starts_with? "0"
return f
end
end
rescue
return f
end
Integer(f.encode(CSV::ConverterEncoding)) rescue f
}
CSV::Converters[:non_octal_float] = lambda { |f|
begin
e = f.encode(CSV::ConverterEncoding)
unless e.starts_with? "0."
if e.starts_with? "0"
return f
end
end
rescue
return f
end
Float(f.encode(CSV::ConverterEncoding)) rescue f
}
CSV::Converters[:string_to_bool] =->(f) {
['true', 'false'].include?(f) ? f=='true' : f
}
@csv_options = {:converters => [:non_octal_integer, :non_octal_float, :string_to_bool]}
@data_type = nil
@udc_id = nil
@filename_suffix = nil
super
end
def expected_output_filename
return "#{@udc_id}_#{@data_type}#{@filename_suffix}.csv"
end
def read_expected_output_file_from_directory! directory
if directory.is_a? String
p "Directory/File:::: #{File.join(directory, self.expected_output_filename)}"
p "@csv_options::: #{@csv_options}"
csv_data = CSV.read File.join(directory, self.expected_output_filename), @csv_options
else
csv_data = CSV.new(directory, @csv_options).read
end
csv_data[0].map! { |h| h.to_sym }
self.clear
self.replace csv_data
end
I am calling read_expected_output_file_from_directory method in my jRuby script like:
expected_dataset.read_expected_output_file_from_directory! File.join($session.script_dir, '..', 'expectedresults')
It is throwing an error:
["wrong number of arguments (given 2, expected 1) (ArgumentError)"]
/home/eip/jruby-9.4.5.0/lib/ruby/stdlib/csv.rb:1821:in `read'
Any help is appreciated ? Thanks
What worked in this as suggested by @Stefan & @Ilmari Karonen :
def read_expected_output_file_from_directory! directory
if directory.is_a? String
p "Directory/File:::: #{File.join(directory, self.expected_output_filename)}"
p "@csv_options::: #{@csv_options}"
csv_data = CSV.read File.join(directory, self.expected_output_filename), **@csv_options**
else
csv_data = CSV.new(directory, **@csv_options**).read
end
csv_data[0].map! { |h| h.to_sym }
self.clear
self.replace csv_data
end
As Stefan notes in the comments above, CSV.read
and CSV.new
expect the options to be passed as keyword arguments.
You can convert a hash to keyword arguments via **
, e.g.:
CSV.new(directory, **@csv_options)
In Ruby 2 (and JRuby 9.3) you could pass a hash as the last argument to a method expecting keyword arguments and it would get implicitly converted into keyword arguments. However, in Ruby 3 (and JRuby 9.4) this no longer happens unless you request it explicitly using **
.