httr CSV 内容读取为整数而不是双精度

Question

我正在尝试使用 httr 从在线来源导入一个 csv，一切都很好，除了它读取一列作为整数时它应该是双精度导致这些值显示为 NA

我正在使用并遇到以下问题。

getdata <- GET(paste("https://rest.zuora.com/v1/files/",r$FileId, sep = ''), auth) invoice <- content(getdata, type = "text/csv")

  Parsed with column specification:
cols(
  Account.external_id__c = col_integer(),
  Invoice.Amount = col_double(),
  Invoice.Balance = col_integer(),
  Invoice.CreatedDate = col_datetime(format = "")
)
Warning: 171 parsing failures.
 row             col               expected actual
2475 Invoice.Balance no trailing characters    .4 
2726 Invoice.Balance no trailing characters    .71
3197 Invoice.Balance no trailing characters    .3 
3287 Invoice.Balance no trailing characters    .5 
3350 Invoice.Balance no trailing characters    .1 
.... ............... ...................... ......
See problems(...) for more details.

任何帮助表示赞赏，谢谢。

Answer 1

该httr::content函数readr::read_csv在您通过时使用type = "text/csv"。您可以将参数传递到read_csv内部content，幸运的是read_csv，您可以定义要导入的 csv 的列类型。

invoice <- content(getdata, type = "text/csv", col_types = cols(
  Account.external_id__c = col_integer(),
  Invoice.Amount = col_double(),
  Invoice.Balance = col_double(),
  Invoice.CreatedDate = col_datetime(format = "")
))

注意Invoice.Balance = col_double()。

查看vignette("column-types", package = "readr")更多信息。