CODEWITHSUNDEEP
curlsession-cookiesrcurlrvesthttr
Frederick Solt
Frederick Solt

Reputation: 376

How to set the domain of cookies in R using rvest, httr, curl?

How can one set the domains for cookies in R using the rvest/httr/curl packages?

I know that one typically sets cookies using a named character vector of names and values (as ?set_cookies reports). Is it possible to specify the domains as well? Or, maybe better yet, to send the whole seven-column Netscape spec that one gets back from httr::cookies()?

The problem arises because I'm trying to automate the shibboleth-based login process for the UK Data Service, which uses a complex set of session cookies (see my earlier question). In the final step of the login process, the site passes two cookies with the same name (JSESSIONID) but for different domains (#HttpOnly_wayf.ukfederation.org.uk and shib.data-archive.ac.uk). As just a named vector, though, the two JSESSIONIDs look alike. I've tried passing back just one or the other, but perhaps not surprisingly it seems that both are necessary.

Upvotes: 0

Views: 797

Answers (1)

Frederick Solt
Frederick Solt

Reputation: 376

Adding the argument config = config(cookiejar = 'cookies.txt') to your rvest command, e.g., submit_form(session = s, form = f, config = config(cookiejar = 'cookies.txt')), does the trick. There's no need to have previously generated a file named cookies.txt, btw: it's all done automagically.

Upvotes: 1

Related Questions