Question

I recently learned that you can use rescue on a line of code in case something goes wrong on that line (see http://www.rubyinside.com/21-ruby-tricks-902.html Tip #21). I have some code that used to look like this:

if obj['key'] && obj['key']['key2'] && obj['key']['key2']['name']
  name = obj['key']['key2']['name']
else
  name = ''
end

With the rescue method, I believe I can change that code into something like this:

name = obj['key']['key2']['name'] rescue ''

If a nil exception is thrown at any level of accessing the hash, it should get caught by the rescue and give me '', which is what I want. I could also choose to set name to nil if that were the desired behavior.

Is there any known danger in doing this? I ask because this seems too good to be true. I have so much ugly code that I'd love to get rid of that looks like the first code example.

Was it helpful?

Solution

Reads good! But it will hit your performance. In my experience rescue is much slower when triggered and slightly slower when it's not. In all cases the if is faster. Other thing to consider, is that exceptions shouldn't be expected and you kind of are with this code. Having a hash so deeply nested might be a good smell that a refactoring is nede

OTHER TIPS

This specific example can now be achieved with Ruby 2.3's dig method.

name = obj.dig 'key', 'key2', 'name'

This will safely access obj['key']['key2']['name'], returning nil if any step fails.

(In general, it's usually advised to use exceptions only for real, unanticipated, errors, though it's understandable in an example like this if the syntax makes it cumbersome.)

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top