Question

I’m trying to extract the second link under the description tag. I have written the following code, but it looks really messy with freads and substrings (just to get it to work). Is there any cleaner way to accomplish this?

magic(Url)->

Tag = “.xml”,

inets:start(),

{ ok, {Status, Headers, Body }} = httpc:request(Url ++ Tag),

{ Xml, Rest } = xmerl_scan:string(Body),

{xmlObj , string , A } = xmerl_xpath:string(“substring-after(substring-after(substring->before(//channel/item/description[1], ‘\”>[link]’) , ‘br’) , ‘href=’)”, Xml),

{ok,_,B} = io_lib:fread(“~6s” , A),

string:sub_string(B,1,string:len(B)-1).

Answer 1

Not a perfect solution, but you may use such xpaths
//channel/item/description[1]/text()[16] and //channel/item/description[1]/text()[24]

extracted strings contains urls + quotes at the beginning, so you may use list matching syntax to cut off quotation marks: [_|Url] = ...

So use this: [{_,_,_,_,[_|U1],_}] = xmerl_xpath:string("//channel/item/description[1]/text()[16]", Xml). to bind U1 with first url.

Test in shell:

11> [{_,_,_,_,[_|U1],_}] = xmerl_xpath:string("//channel/item/description[1]/text()[16]", Xml). 
[{xmlText,[{description,5},{item,5},{channel,1},{rss,1}],
          16,[],"\"http://www.reddit.com/user/escaped_reddit",text}]
12> 
12> U1.
"http://www.reddit.com/user/escaped_reddit"
13> 
13> 
13> [{_,_,_,_,[_|U2],_}] = xmerl_xpath:string("//channel/item/description[1]/text()[24]", Xml). 
[{xmlText,[{description,5},{item,5},{channel,1},{rss,1}],
          24,[],
          "\"http://www.reddit.com/r/erlang/comments/y62wf/how_to_use_ranch/",
          text}]
14> 
14> U2.
"http://www.reddit.com/r/erlang/comments/y62wf/how_to_use_ranch/"