# iterate over an array and delete elements conditionally

The following logic:

Iterate over the URLs array and If the HTTP request with the given URL does not timeout, continue. If the URL times out, remove it from the array and continue with the next one until the array is exhausted.

Has been implemented in the following way:

@urls.delete_if do |url|
begin
doc = perform_request(some_params)
break
rescue TimeoutError
Rails.logger.warn("URL #{url} times out, will be removed from list")
true
end
end


Anyone for a cleaner solution? Basically, I want to iterate over an array and remove elements from it if there is a timeout, but if the URL works loop should end and app should continue processing.

-
Specifically, hate the break in there - maybe someone can propose a cleaner solution. –  Emo Jan 4 '13 at 15:23
Btw "true" in the rescue can be removed, not needed. –  Emo Jan 4 '13 at 15:33
if not needed just edit the question! I have a doubt, why do you want to modify inplace @urls? wouldn't do a functional solution? –  tokland Jan 4 '13 at 17:57

note : your break is useless, because if the block returns a falsey value, the url is kept.

Maybe it can just be a bit more expressive ?

def timeouts?( url )
perform_request( url ) && false
rescue TimeoutError
true
end

@urls.delete_if {|url| timeouts?( url ) }


You could even make timeouts? a method on your @urls objects, for extra readability :

@urls.delete_if( &:timeouts? )


if you want to keep @tokland's functionnal approach, just use :

@kept_urls = @urls.reject{|url| timeouts?( url ) }


however, it would not be fully "functionnal", because timeouts may not return the same results with the same arguments, depending on the execution environment.

-

You may use drop_while to get rid of iterative operator break:

@urls = @urls.drop_while do |url|
begin
doc = perform_request(some_params)
false
rescue TimeoutError
Rails.logger.warn("URL #{url} times out, will be removed from list")
true
end
end

-
the problem is doc won't be visible outside the block's scope. –  tokland Jan 4 '13 at 21:23

I'd take a functional approach (no in-place updates):

url, doc = @urls.map_detect do |url|
begin
[url, perform_request(some_params)]
rescue TimeoutError
Rails.logger.warn("URL #{url} timed out")
false
end
end


You can then use url to update @urls, but it's better not to mix the algorithm (even if it's so simple) with the updates required by the app.