ansaurus

Question

Answer 1

+3 A:

>> doc = Hpricot.parse('<html><img src="foo" class="int" /><img src="bar" bar="42" /><img src="foobar" class="int"></html>')
=> #<Hpricot::Doc {elem <html> {emptyelem <img class="int" src="foo">} {emptyelem <img src="bar" bar="42">} {emptyelem <img class="int" src="foobar">} </html>}>
>> doc.search("img")[1][:bar]
=> "42"
>> doc.search("img") - doc.search("img.int")
=> [{emptyelem img src"bar" bar"42"}]

Once you have results from search you can use normal array operations. nonstandard attributes are accessible through [].

Ben Hughes 2009-06-29 14:52:41

wow, pretty easy, so I can use "collection_one - collection_two" to remove all elements from collection_one which are in collection_two? Thanks!

ole_berlin 2009-06-29 15:01:20

Important note, after `z = x - y`, x won't change, z will just contain all of x that isn't in y.

rampion 2009-06-29 15:09:18

yep, and + to add collections, e.g. `all_images - (internal_images + external_images)`.

Ben Hughes 2009-06-29 15:09:32

Answer 2

+2 A:

Check out the not CSS selector.

(hbody."img:not(.int)")
(hbody."img:not(.ext)")

Unfortunately, it doesn't seem you can concat not expressions. You might want to fetch all img nodes and remove those where the .css selector doesn't include neither .int nor .ext. Additionally, you could use the difference operator to calculate which elements are not part of both collections.

Use the .remove method to remove nodes or elements: Hpricot Altering documentation.

Simone Carletti 2009-06-29 15:02:44

I could finally achieve what I wanted by chaining the searches: hbody.search("img").search(":not(.int)").search(":not(.ext)") returns all images where the class is not "int" and not "ext". After removing them I can check for the other attributes and do basically the same. Thanks again!

ole_berlin 2009-06-29 17:04:21

ansaurus

tags:

views:

answers:

Hpricot Element intersection

related questions