在ruby中将.json转换为.csv

时间:2011-10-21 04:34:23

标签: ruby ruby-on-rails-3 json csv

我想使用ruby将.json文件转换为.csv文件。请帮助我这样做。

还提出任何工具来实现这一目标。

7 个答案:

答案 0 :(得分:24)

尝试这样的事情:

require 'csv'
require 'json'

csv_string = CSV.generate do |csv|
  JSON.parse(File.open("foo.json").read).each do |hash|
    csv << hash.values
  end
end

puts csv_string

答案 1 :(得分:9)

实际写入文件...

require 'csv'
require 'json'

CSV.open("your_csv.csv", "w") do |csv| #open new file for write
  JSON.parse(File.open("your_json.json").read).each do |hash| #open json to parse
    csv << hash.values #write value to file
  end
end

答案 2 :(得分:5)

我认为在ruby中将JSON转换为CSV的简单方法是使用json2csv ruby​​ gem。

PS:我可能有偏见,因为我是这个人的作者。

json2csv github repo

运行命令

json2csv convert data/sample.json

输入文件

在这种特殊情况下,它将转换为以下json:

cat data/sample.json

{
"12345": {
    "Firstname": "Joe",
    "Lastname": "Doe",
    "Address": {
        "Street": "#2140 Taylor Street, 94133",
        "City": "San Francisco",
        "Details": {
            "note": "Pool available"
        }
    }
},

"45678": {
    "Firstname": "Jack",
    "Lastname": "Plumber",
    "Address": {
        "Street": "#111 Sutter St, 94104",
        "City": "San Francisco",
        "Details": {
            "note": "Korean Deli near to main entrance"
        }
    }
}

<强>输出

cat data/sample.json.csv

id,Firstname,Lastname,Address.Street,Address.City,Address.Details.note
12345,Joe,Doe,"#2140 Taylor Street, 94133",San Francisco,Pool available
45678,Jack,Plumber,"#111 Sutter St, 94104",San Francisco,Korean Deli near to main entrance

答案 3 :(得分:4)

基于@Alex's answer,但添加 csv标头示例测试

# utils.rb
require "csv"
module Utils
  def self.array_of_hashes_to_csv(array_of_hashes)
    CSV.generate do |csv|
      csv << array_of_hashes.first.keys
      array_of_hashes.each { |hash| csv << hash.values }
    end
  end
end

# utils_test.rb
class UtilsTest < MiniTest::Unit::TestCase
  def test_array_of_hashes_to_csv
    array_of_hashes = [
      { :key1 => "value1", :key2 => "value2" },
      { :key1 => "value3", :key2 => "value4" }
    ]
    expected_result = "key1,key2\nvalue1,value2\nvalue3,value4\n"
    assert_equal(expected_result, Utils.array_of_hashes_to_csv(array_of_hashes))
  end
end

答案 4 :(得分:3)

修改:

下面描述的此功能现在以gem提供。使用gem install json_converter安装后,可以使用以下代码段从有效的JSON字符串或对象生成CSV:

require 'json_converter'
json_converter= JsonConverter.new

# Assume json is a valid JSON string or object
csv = json_converter.generate_csv json

原始答案:

如果您的JSON数据相对简单(没有嵌套或数组),Alex的答案可能是解决此问题的最简洁方法。

但是,如果你需要考虑数组和嵌套对象,我试图将web version这样的转换器移植到ruby。它可以找到here。处理数据实际重组的方法是array_fromflatten

array_from方法尝试识别&#34;行&#34;对于给定的数据集,数据看起来像。它并不完美,您可能希望针对不同的数据集调整此部分。

# Attempt to identify what a "row" should look like
def array_from(json)
  queue, next_item = [], json
  while !next_item.nil?

    return next_item if next_item.is_a? Array

    if next_item.is_a? Hash
      next_item.each do |k, v|
        queue.push next_item[k]
      end
    end

    next_item = queue.shift
  end

  return [json]
end

flatten方法以递归方式迭代JSON对象,并生成表示标题和值的对象。如果对象是嵌套的,则其列的标题将以其父键为前缀,由/字符分隔。

# The path argument is used to construct header columns for nested elements
def flatten(object, path='')
  scalars = [String, Integer, Fixnum, FalseClass, TrueClass]
  columns = {}

  if [Hash, Array].include? object.class
    object.each do |k, v|
      new_columns = flatten(v, "#{path}#{k}/") if object.class == Hash
      new_columns = flatten(k, "#{path}#{k}/") if object.class == Array
      columns = columns.merge new_columns
    end

    return columns
  elsif scalars.include? object.class
      # Remove trailing slash from path
      end_path = path[0, path.length - 1]
      columns[end_path] = object
      return columns
  else
    return {}
  end
end

如果原始JSON中有任何null值,则在尝试转换之前,您需要将这些值转换为nil以外的其他值 - 您通常会以如果你不这样,那么行不均匀。 nils_to_strings方法处理:

# Recursively convert all nil values of a hash to empty strings
def nils_to_strings(hash)
  hash.each_with_object({}) do |(k,v), object|
    case v
    when Hash
      object[k] = nils_to_strings v
    when nil
      object[k] = ''
    else
      object[k] = v
    end
  end
end

这是一个如何使用它的简短示例:

json = JSON.parse(File.open('in.json').read)
in_array = array_from json
in_array.map! { |x| nils_to_strings x }

out_array = []
in_array.each do |row|
  out_array[out_array.length] = flatten row
end

headers_written = false
CSV.open('out.csv', 'w') do |csv|
  out_array.each do |row|
    csv << row.keys && headers_written = true if headers_written === false
    csv << row.values
  end
end

最后,这里有一些输入/输出示例:

输入:

{
  "Forms": [
    {
      "Form": {
        "id": "x",
        "version_id": "x",
        "name": "x",
        "category": "",
        "subcategory": null,
        "is_template": null,
        "moderation_status": "x",
        "display_status": "x",
        "use_ssl": "x",
        "modified": "x",
        "Aggregate_metadata": {
          "id": "x",
          "response_count": "x",
          "submitted_count": "x",
          "saved_count": "x",
          "unread_count": "x",
          "dropout_rate": "x",
          "average_completion_time": null,
          "is_uptodate": "x"
        }
      },
      "User": {
        "username": "somedude@example.com"
      }
    },
    {
      "Form": {
        "id": "x",
        "version_id": "x",
        "name": "x",
        "category": "",
        "subcategory": null,
        "is_template": null,
        "moderation_status": "x",
        "display_status": "x",
        "use_ssl": "x",
        "modified": "x",
        "Aggregate_metadata": {
          "id": "x",
          "response_count": "x",
          "submitted_count": "x",
          "saved_count": "x",
          "unread_count": "x",
          "dropout_rate": "x",
          "average_completion_time": null,
          "is_uptodate": "x"
        }
      },
      "User": {
        "username": "somedude@example.com"
      }
    }
  ]
}

输出:

Form/id,Form/version_id,Form/name,Form/category,Form/subcategory,Form/is_template,Form/moderation_status,Form/display_status,Form/use_ssl,Form/modified,Form/Aggregate_metadata/id,Form/Aggregate_metadata/response_count,Form/Aggregate_metadata/submitted_count,Form/Aggregate_metadata/saved_count,Form/Aggregate_metadata/unread_count,Form/Aggregate_metadata/dropout_rate,Form/Aggregate_metadata/average_completion_time,Form/Aggregate_metadata/is_uptodate,User/username
x,x,x,"","","",x,x,x,x,x,x,x,x,x,x,"",x,somedude@example.com
x,x,x,"","","",x,x,x,x,x,x,x,x,x,x,"",x,somedude@example.com

答案 5 :(得分:1)

要从json转换为csv的Ruby代码。

通过将数组转换为双引号字符串来处理数组。

#json-to-csv.rb
require 'json'
require 'csv'

file = File.read('./input.json')
hash = JSON.parse(file)

CSV.open('./output.csv', 'w') do |csv|
  headers = hash.first.keys
  csv << headers

  hash.each do |item|
    values = item.values
    printable_values = Array.new
    values.each do |value|

      printable_values << value.to_s.gsub(/\[|\]/,'').gsub(/"/,'\'')

    end

    csv << printable_values

  end

end

答案 6 :(得分:0)

这是句柄标题&amp;嵌套的json。

require 'csv'
require 'json'

@headers = []
file = File.open('file.json')
JSON.parse(file.read).each do |h|
  h.keys.each do |key|
    @headers << key
  end
end

uniq_headers = @headers.uniq
file = File.open('file.json')
finalrow = []

JSON.parse(file.read).each do |h|
  final = {}
  @headers.each do |key2|
    final[key2] = h[key2]
  end

  finalrow << final
end

CSV.open('output.csv' , 'w') do |csv|
  csv << uniq_headers
  finalrow.each do |deal|
    csv << deal.values
  end
end