Home/ Questions/Q 7580295
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-30T17:52:54+00:00

I’m fairly new to ruby. Recently, I wanted to extract a portion of a

  • 0

I’m fairly new to ruby. Recently, I wanted to extract a portion of a string from the n’th character of said string to the end.

Doing something like s[n,(s.size – n)] seemed pretty inelegant to me, so I asked a couple of friends.

One suggested I try s[n..-1], and sure enough that works, but he couldn’t give me a good reason why it should work. I find the fact that it works rather puzzling, as the following output from irb1.9 should explain:


> s = "0123456789"
=> "0123456789"
> s[2..-1]
=> "23456789"
> (2..-1).to_a
=> []

So as you can see, the range object 2..-1 is empty — it has no members, which is absolutely what you expect if you go upwards in value from 2 to -1. This is consistent with the documentation for how range objects should work.

The documentation for indexing a string with a range clearly says: “If given a range, a substring containing characters at offsets given by the range is returned” — but that is an empty set.

I also can find no examples in “The Ruby Programming Language” or in the Ruby docs in which a string is indexed using s[n..-1] or the like, and can find no examples of it in other official sources. It appears to be folklore, however, that it works even though nothing in the manuals indicate that you can index a string with a range this way, and get the result you get even though the range has no members.

Yet, my friend was correct, it works.

So, could someone please explain why this works to me? I’m also very much interested in knowing if the fact that it works is a fluke of MRI/YARV or if this is absolutely expected to work in all Ruby implementations, and if so, where is it documented to work?

EDITED TO ADD:

An answerer below claimed that only the range’s begin and end attributes matter for these purposes, but I can find no documentation of that in TRPL or in the Ruby documentation. The answer also claims that there are indeed examples of such “mixed-sign” range indexing, but the only one I could find was in a context where the mixed-range index was shown to produce nil, not a slice of a string. I therefore don’t find that answer satisfying.

EDITED TO ADD:

It appears that the correct answer is that this is indeed a defect in the Ruby documentation.

EDITED TO ADD:

The bug was fixed by the Ruby documentation team: see https://bugs.ruby-lang.org/issues/6106

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    This is a bug in the documentation.

    Ruby’s documentation has sucked since the Pickaxe book descended like a meteor on matz’s actually correct and comprehensive HTML doc. This is a subject that still irritates me on occasion. The answer to your question, from 1.4: link

    self[nth]

    Retrieves the nth item from an array. Index starts from zero. If index is the negative, counts backward from the end of the array. The index of the last element is -1. Returns nil, if the nth element is not exist in the array.

    self[start..end]

    Returns an array containing the objects from start to end, including both ends. If … is used (instead of ..), then end is not included. if end is larger than the length of the array, it will be rounded to the length. If start is out of an array range , returns nil. And if start is larger than end with in array range, returns empty array ([]).

    -1 is the last index of an array by definition, as a convenience.

    • 0