Can someone explain the difference: Array[] and Array.fetch

array = [“a”,“b”,“c”] # => [“a”, “b”, “c”]
array[1] # => “b”
array.fetch 1 # => “b”

==================================================

So far, so fine.

==================================================

array[666] # => nil
array.fetch 666

IndexError: index 666 out of array
from (irb):5:in `fetch’
from (irb):5

This may seem like a trival difference between the two, but
I am wondering in general why they behave differently.
Or in other words, why array[] does not return an error
whereas .fetch specifically does.

Marc H. wrote:

array[666] # => nil
whereas .fetch specifically does.
You might say that’s the purpose of fetch, to raise an exception for
out-of-bounds indexes when [] just returns nil. Pick the method you want
depending on the error handling you want.

Marc H. [email protected] wrote:

array[666] # => nil
whereas .fetch specifically does.
fetch is more general; if you want it to behave like in this case,
say array.fetch(666,nil). m.

On 06.06.2009 03:03, Marc H. wrote:

array[666] # => nil
whereas .fetch specifically does.
I am not the author of that class so I can only speculate. The reason
is to have a proper tool for different situations. Often you do not
care whether the array index does exist. For example, with [] you can
do

x = []
x[10] = 123

Without having to worry about array dimensions.

If on the other hand you want to see an error if an index out of the
current range is used or want to provide a default value you can use
#fetch. Btw, these methods are defined with similar semantics for Hash.

Kind regards

robert

Hi,

I too learned a lot from that thread about this. Since I’m still new to
ruby I’ve to use the docs [1] nearly daily. I wish this kind of “hint”
would be added to both methods …

  • Markus

[1] e.g. class Array - RDoc Documentation

Thanks to all for the explanation!