2016-12-28 59 views
0

我试图通过RCurl从国家信息中心下载pdf,但我遇到了一些麻烦。对于这个例子URL,我希望pdf对应于默认设置,除了“报告格式”应该是“PDF”。当我运行下面的脚本时,它会保存与选择其他按钮相关的文件(“...的父母”/ HMDA--不是默认的)。我试着将这些输入元素添加到params,但它没有改变任何东西。有人能帮我找出问题吗?谢谢。将正确的参数传递给RCurl/postForm

library(RCurl) 
curl = getCurlHandle() 
curlSetOpt(cookiejar = 'cookies.txt', curl = curl) 
params = list(rbRptFormatPDF = 'rbRptFormatPDF') 

url = 'https://www.ffiec.gov/nicpubweb/nicweb/OrgHierarchySearchForm.aspx?parID_RSSD=2162966&parDT_END=99991231' 
html = getURL(url, curl = curl) 
viewstate = sub('.*id="__VIEWSTATE" value="([0-9a-zA-Z+/=]*).*', '\\1', html) 
event = sub('.*id="__EVENTVALIDATION" value="([0-9a-zA-Z+/=]*).*', '\\1', html) 
params[['__VIEWSTATE']] = viewstate 
params[['__EVENTVALIDATION']] = event 
params[['btnSubmit']] = 'Submit' 
result = postForm(url, .params=params, curl=curl, style='POST') 

writeBin(as.vector(result), 'test.pdf') 

回答

1

这是否提供正确的PDF?

library(httr) 
library(rvest) 
library(purrr) 

# setup inane sharepoint viewstate parameters 
res <- GET(url = "https://www.ffiec.gov/nicpubweb/nicweb/OrgHierarchySearchForm.aspx", 
      query=list(parID_RSSD=2162966, parDT_END=99991231)) 

# extract them 
pg <- content(res, as="parsed") 
hidden <- html_nodes(pg, xpath=".//form/input[@type='hidden']") 
params <- setNames(as.list(xml_attr(hidden, "value")), xml_attr(hidden, "name")) 

# pile on more params 
params <- c(
    params, 
    grpInstitution = "rbCurInst", 
    lbTopHolders = "2961897", 
    grpHMDA = "rbNonHMDA", 
    lbTypeOfInstitution = "-99", 
    txtAsOfDate = "12/28/2016", 
    txtAsOfDateErrMsg = "", 
    lbHMDAYear = "2015", 
    grpRptFormat = "rbRptFormatPDF", 
    btnSubmit = "Submit" 
) 

# submit the req and save to disk 
POST(url = "https://www.ffiec.gov/nicpubweb/nicweb/OrgHierarchySearchForm.aspx", 
    query=list(parID_RSSD=2162966, parDT_END=99991231), 
    add_headers(Origin = "https://www.ffiec.gov"), 
    body = params, 
    encode = "form", 
    write_disk("/tmp/output.pdf")) -> res2 
+0

是的,这个作品 - 谢谢!我想你确实需要明确地指定所有参数,即使它们显示为默认值 – sirallen

+0

似乎并不需要“lbTopHolders”和“lbHMDAyear” – sirallen

+0

隐藏的“API”确实是可怕的;-) – hrbrmstr