Range comparisons

ruby-talk
1

Something came up as I was testing some ideas. I know Range objects
can indicate whether a specific object is an included element, but
something I thought might work didn't:

irb(main):001:0> (1..10).include?(2..6)
=> false

I know a simple case like this can be covered with (1..10).include? 2
and (1..10).include? 6 (range1 includes both range2.first and
range2.last), but doing it that way seems not entirely right,
especially when it comes to other cases. A big problem, as I see it,
would be trying to figure out what to do if the second range excludes
its end.

Maybe I'm just thinking about this completely wrong. Maybe I should
be using Sets or something. Knowing about intersects would be nice,
too.

Any thoughts?

···

--
-yossef

2

Here is an interesting discussion.

http://opensoul.org/2007/2/13/ranges-include-or-overlap-with-ranges

Anyway, nothing that would require more than 10 lines of code.

3

Similar discussions have come up here before. The net was, that because of all sorts of issues (which you partly identified yourself already) it would not be wise to extend functionality of Range#include? beyond the current behavior. You'll sure find more detail in the archives but one reason for example is that other Enumerable's include? methods always check for single element membership only. For all the other operations there is - and & etc.

18:13:10 [contentreporter]: irb -r set
irb(main):001:0> s1=(1..10).to_set
=> #<Set: {5, 6, 1, 7, 2, 8, 3, 9, 4, 10}>
irb(main):002:0> s2=(1...10).to_set
=> #<Set: {5, 6, 1, 7, 2, 8, 3, 9, 4}>
irb(main):003:0> s1 == s2
=> false
irb(main):004:0> s1 - s2
=> #<Set: {10}>
irb(main):005:0> s2 - s1
=> #<Set: {}>
irb(main):010:0> s2.superset? s1
=> false
irb(main):011:0> s1.superset? s2
=> true

Kind regards

  robert

···

On 04.06.2007 17:56, Yossef Mendelssohn wrote:

Something came up as I was testing some ideas. I know Range objects
can indicate whether a specific object is an included element, but
something I thought might work didn't:

irb(main):001:0> (1..10).include?(2..6)
=> false

I know a simple case like this can be covered with (1..10).include? 2
and (1..10).include? 6 (range1 includes both range2.first and
range2.last), but doing it that way seems not entirely right,
especially when it comes to other cases. A big problem, as I see it,
would be trying to figure out what to do if the second range excludes
its end.

Maybe I'm just thinking about this completely wrong. Maybe I should
be using Sets or something. Knowing about intersects would be nice,
too.

Any thoughts?

4

Thanks to both of you (CHubas and Robert). I didn't see the replies
because either I don't get replies to my threads e-mailed to me or I'm
just oblivious.

I've written a few things myself, things on the order of
Range#contains?, Range#contained_by?, Range#overlaps?. The real
trick, however comes when I'm using extensions built on top of Range,
extensions that allow a range-ish objects that have open ends. I'm
not talking about excluding the end, like 1...10, but allowing the
absence of an end, like 1..nil. Now, that can be done easily enough
with numbers and Infinity (1.0/0), but it gets troublesome when using,
say, Times or Dates.

···

--
-yossef