From: Rob Biedenharn <Rob@...>
Date: 2008-08-19T00:45:31+09:00
Subject: Re: how to move on to next url if current one is a dud


On Aug 18, 2008, at 10:40 AM, Stefano Crocco wrote:

> On Monday 18 August 2008, Adam Akhtar wrote:
>> hmm wasnt too sure how to make the subject both concise and
>> understandable at the same time so my apologies if i was vague.
>>
>> I searched under "ignore missing" but didnt come up wiht anything.
>> Perhaps someone can help me.
>>
>> Im using open-uri to fetch the html from various ebay webpages.
>> Basically the results for certain categories. I have all the relevant
>> urls tucked into an array of hashes eg. Computers =>
>> www.ebay.blah.blah.blah
>>
>> I then cycle through the array feeding open uri the urls and doing my
>> biz on the html one by one.
>>
>> However sometimes there are no products for a certain category for  
>> that
>> given day (im focussing on products located in japan so the number of
>> results is drastically reduced) so when open uri complains like this
>>
>> ebay.rb:71:in `read': No such file or directory -
>> http://research.ebay.co.uk/ws/
>> search/GetSaleAnalysis? 
>> sojs=0&salic=104&sacqyop=9&sacqy=1000&sacat=13851&ft
>> rv=17
>> &fis 
>> =2&fts=-1&fn=4&frts=0&sc=AvgSoldPrice&showdiag=0&soexpand=1&soraw=0
>> (Errno::
>> ENOENT)
>>
>> Im fine with it complaining but it stops the programme. How do i  
>> make it
>> simply skip/ignore such urls and move on to the next one in the hash?
>>
>> any help greatly appreciated.
>
> Wrapping the call to open in a begin/rescue pair should do what you  
> want:
>
> begin open( url )
> rescue SystemCallError
> end
>
> I hope this helps
>
> Stefano


You might want to rescue just Errno::ENOENT.

I often do:

     rescue Timeout::Error, OpenURI::HTTPError, Errno::ENOENT

for this kind of thing.

-Rob

Rob Biedenharn		http://agileconsultingllc.com
Rob@AgileConsultingLLC.com