Submit POST form when rvest doesn't recognize submit button

Question

I would like to submit the following form (the form appears after you click on link "Kliknite na ..."): http://www1.biznet.hr/HgkWeb/do/extlogon

I have to enter one parameter, named "OIB" and submit the form by clicking "Trazi".

Here is my code:

library(httr)
library(rvest)

sess <- html_session("http://www1.biznet.hr/HgkWeb/do/extlogon")
search_page <- sess %>%
  follow_link(1)
form <- html_form(search_page)[[6]]
fill_form <- set_values(form, 'clanica.cla_oib' = '94989605030')
firma_i <- submit_form(search_page, fill_form, submit = 'submit')

Last line produces an error:

Error: Unknown submission name 'submit'. Possible values: clanica.asTextDatumGasenjaTo, clanica.asTextUdr_id

I don't understand why rvest recognize this two parameters as submit buttons when they don't contain submit name or type. And why rvest doesn't recognize submit button "Trazi" as submit parameter? In, short, how to change filled form to execute the form?

score 7 · Accepted Answer · answered Aug 27 '18 at 15:30

The problem is that some of the input miss the type attr, and rvest does not check this appropriately.

To illustrate the problem:

library(httr)
library(rvest)
#> Loading required package: xml2

sess <- html_session("http://www1.biznet.hr/HgkWeb/do/extlogon")
search_page <- sess %>%
  follow_link(1)
#> Navigating to /HgkWeb/do/extlogon;jsessionid=88295900F3F932C85A25BB18F326BE28
form <- html_form(search_page)[[6]]
fill_form <- set_values(form, 'clanica.cla_oib' = '94989605030')

Some of the fields do not have the type attribute:

sapply(fill_form$fields, function(x) '['(x, 'type'))
#> $clanica.limitSearchToActiveCompany.type
#> [1] "radio"
#> 
#> $clanica.limitSearchToActiveCompany.type
#> [1] "radio"
#> 
#> $joinBy.useInnerJoin.type
#> [1] "checkbox"
#> 
#> $nazivTvrtke.type
#> [1] "text"
#> 
#> $nazivZapocinjeSaPredanomVrijednoscu.type
#> [1] "checkbox"
#> 
#> $clanica.cla_jmbp.type
#> [1] "text"
#> 
#> $clanica.cla_mbs.type
#> [1] "text"
#> 
#> $clanica.cla_oib.type
#> [1] "text"
#> 
#> $asTextKomoraId.NA
#> NULL
#> 
#> $clanica.asTextOpc_id.NA
#> NULL
#> 
#> $clanica.cla_opcina.type
#> [1] "hidden"
#> 
#> $clanica.asTextNas_id.NA
#> NULL
#> 
#> $clanica.cla_naselje.type
#> [1] "hidden"
#> 
#> $clanica.pos_id.NA
#> NULL
#> 
#> $clanica.postaNaziv.type
#> [1] "hidden"
#> 
#> $clanica.cla_ulica.type
#> [1] "text"
#> 
#> $clanica.asTextDatumUpisaFrom.type
#> [1] "text"
#> 
#> $clanica.asTextDatumUpisaTo.type
#> [1] "text"
#> 
#> $clanica.asTextDatumGasenjaFrom.type
#> [1] "text"
#> 
#> $clanica.asTextDatumGasenjaTo.type
#> [1] "text"
#> 
#> $clanica.asTextUdr_id.NA
#> NULL
#> 
#> $clanica.asTextVel_id.NA
#> NULL
#> 
#> $nkd2007.type
#> [1] "text"
#> 
#> $nkd2007PretrazivanjePoGlavnojDjelatnosti.type
#> [1] "radio"
#> 
#> $nkd2007PretrazivanjePoGlavnojDjelatnosti.type
#> [1] "radio"
#> 
#> $submit.type
#> [1] "submit"
#> 
#> $org.apache.struts.taglib.html.CANCEL.type
#> [1] "submit"
#> 
#> $orderBy.order1.NA
#> NULL
#> 
#> $orderBy.order2.NA
#> NULL
#> 
#> $limit.type
#> [1] "text"
#> 
#> $searchForRowCount.type
#> [1] "checkbox"
#> 
#> $joinBy.gfiGodina.NA
#> NULL
#> 
#> $joinBy.gfiBrojZaposlenihFrom.type
#> [1] "text"
#> 
#> $joinBy.gfiBrojZaposlenihTo.type
#> [1] "text"
#> 
#> $joinBy.gfiUkupniPrihodFrom.type
#> [1] "text"
#> 
#> $joinBy.gfiUkupniPrihodTo.type
#> [1] "text"

This messes up the internal function submit_request and specifically the Filter() in it.

It's referenced here, and a fix is proposed in this PR, but it hasn't been merged since Jul 2016, so don't hold your breath.

The fix in the PR basically check if a type attr is present:

  # form.R, row 280
  is_submit <- function(x) 'type' %in% names(x) &&
                           tolower(x$type) %in% c("submit", "image", "button")

For a quick fix you can change the data you have, overriding the NULL attr, with a random type:

fill_form$fields <- lapply(fill_form$fields, function(x) {
  null_type = is.null(x$type)
  if (null_type) x$type = 'text'
  x
})


firma_i <- submit_form(search_page, fill_form, submit = 'submit')
firma_i
#> <session> http://www1.biznet.hr/HgkWeb/do/fullSearchPost
#>   Status: 200
#>   Type:   text/html;charset=UTF-8
#>   Size:   4366

Created on 2018-08-27 by the reprex package (v0.2.0).

Submit POST form when rvest doesn't recognize submit button

1 Answers1