CSV.foreach 不读取 CSV 文件中的第一列

Question

第一次学习 Ruby 自动清理一些 CSV 文件。

我已经设法将下面的脚本从其他 SO 问题拼凑起来，但由于某种原因，脚本没有读取原始 CSV 文件的第一列。如果我添加一个虚拟的第一列，一切正常。我错过了什么？

require 'csv'

COLUMNS = ['SFID','Date','Num','Transaction Type']

CSV.open("invoicesfixed.csv", "wb",
  :write_headers=> true,
  :headers => ["Account__c","Invoice_Date__c","Invoice_Number__c","Transaction_Type__c"]) do |csv|

  CSV.foreach('invoices.csv', :headers=>true, :converters => :all) do |row|


    #convert date format to be compatible with Salesforce
    row['Date'] = Date.strptime(row['Date'], '%m/%d/%y').strftime('%Y-%m-%d')
    csv << COLUMNS.map { |col| row[col] }

  end
end

这个输入文件：

Transaction Type,Date,Num,SFID
Invoice,7/1/19,151466,SFID1
Invoice,7/1/19,151466,SFID2
Invoice,7/1/19,151466,SFID3
Invoice,7/1/19,151466,SFID4
Invoice,7/1/19,151466,SFID5
Invoice,7/1/19,151466,SFID6
Invoice,7/1/19,151153,SFID7
Sales Receipt,7/1/19,149487,SFID8
Sales Receipt,7/1/19,149487,SFID9
Sales Receipt,7/1/19,149758,SFID10
Sales Receipt,7/1/19,149758,SFID11

产生这个输出：

Account__c,Invoice_Date__c,Invoice_Number__c,Transaction_Type__c
SFID1,2019-07-01,151466,
SFID2,2019-07-01,151466,
SFID3,2019-07-01,151466,
SFID4,2019-07-01,151466,
SFID5,2019-07-01,151466,
SFID6,2019-07-01,151466,
SFID7,2019-07-01,151153,
SFID8,2019-07-01,149487,
SFID9,2019-07-01,149487,
SFID10,2019-07-01,149758,
SFID11,2019-07-01,149758,

但是，这个输入：

Dummy,Transaction Type,Date,Num,SFID
,Invoice,7/1/19,151466,SFID1
,Invoice,7/1/19,151466,SFID2
,Invoice,7/1/19,151466,SFID3
,Invoice,7/1/19,151466,SFID4
,Invoice,7/1/19,151466,SFID5
,Invoice,7/1/19,151466,SFID6
,Invoice,7/1/19,151153,SFID7
,Sales Receipt,7/1/19,149487,SFID8
,Sales Receipt,7/1/19,149487,SFID9
,Sales Receipt,7/1/19,149758,SFID10
,Sales Receipt,7/1/19,149758,SFID11

产生正确的输出：

Account__c,Invoice_Date__c,Invoice_Number__c,Transaction_Type__c
SFID1,2019-07-01,151466,Invoice
SFID2,2019-07-01,151466,Invoice
SFID3,2019-07-01,151466,Invoice
SFID4,2019-07-01,151466,Invoice
SFID5,2019-07-01,151466,Invoice
SFID6,2019-07-01,151466,Invoice
SFID7,2019-07-01,151153,Invoice
SFID8,2019-07-01,149487,Sales Receipt
SFID9,2019-07-01,149487,Sales Receipt
SFID10,2019-07-01,149758,Sales Receipt
SFID11,2019-07-01,149758,Sales Receipt

任何想法为什么会发生这种情况？

Answer 1

我有一个类似的问题，虽然运行你的例子有效。我意识到这个问题（至少对我而言）是我正在使用 Excel 中的“另存为 UTF-8 CSV”创建 CSV 文件。

这将BOM添加到文件的开头 - 在第一列标题名称之前，因此 row['firstColumnName'] 返回 nil。

将文件另存为 CSV 为我解决了这个问题。