# What's the ruby way to sort string with cases in mind

Hi,

I’d like to sort an array of strings case-insensitively.
However, if 2 strings are case-insensitively same, I want capital
letters to come first.

a = %w(a c b A C B)
a.sort_by { |i| i.upcase } # => [“a”, “A”, “b”, “B”, “c”, “C”]

Actually, I want [“A”, “a”, “B”, “b”, “C”, “c”] .
I can get the result with the following trick.

a.sort_by {|i| i.upcase + i}

But I think there’s a more rubyish way.
How do you do that?

Sam

Sam K. wrote:

I’d like to sort an array of strings case-insensitively.
Alternate method (slightly “faster”, maybe):
a = %w(a c b A C B)
a.sort {|a,b| a.casecmp b} #=> [“a”, “A”, “b”, “B”, “c”, “C”]

Actually, I want [“A”, “a”, “B”, “b”, “C”, “c”] .
I can get the result with the following trick.

a.sort_by {|i| i.upcase + i}
Clever.

But I think there’s a more rubyish way.
How do you do that?
No clue. Maybe implement your own comparator? {|a,b| dict_order(a,b)}?

Devin

sort routine is not going to be anywhere near as clean as what you’ve
got.

My best try so far is like that:

a.sort{ |i,j| (i.upcase == j.upcase)? (i<=>j):(i.upcase<=>j.upcase) }

Not elegant, but not too ugly (or is it?). I’ve got a feeling there
must be a way to simplify this.

Regards,
Rimantas

Sam K. wrote:

But I think there’s a more rubyish way.
How do you do that?

Hi Sam,

I think that’s probably the most concise and rubyish way. You can’t do
a (simple) sort by ascii ordinal (i[0] in 1.8, or i.ord in 1.9), as
that would give you all the capitals first. And implementing your own
sort routine is not going to be anywhere near as clean as what you’ve
got. So yeah, imo, go with what you have, it’s nice.

Regards,
Jordan

On 20.09.2006 09:19, Rimantas L. wrote:

sort routine is not going to be anywhere near as clean as what you’ve
got.

My best try so far is like that:

a.sort{ |i,j| (i.upcase == j.upcase)? (i<=>j):(i.upcase<=>j.upcase) }

Not elegant, but not too ugly (or is it?). I’ve got a feeling there
must be a way to simplify this.

a.sort {|i,j| i.upcase == j.upcase ? i <=> j : i.upcase <=> j.upcase}

Maybe a bit more efficient:

a.sort do |i,j|
cmp = i.downcase <=> j.downcase
cmp == 0 ? i <=> j : cmp
end

Cheers

``robert``

I shouldn’t have underestimated my fellow rubyists. I guess
implementing your own sort routine can be near as clean.

Regards,
Jordan

Hi,

“Rimantas L.” [email protected] writes:

must be a way to simplify this.
a.sort{ |i,j| [i.upcase, i] <=> [j.upcase, j] }

a.sort_by{ |i| [i.upcase, i] }

a.sort{ |i,j| [i.upcase, i] <=> [j.upcase, j] }

a.sort_by{ |i| [i.upcase, i] }

Great! I guess we have the winner

Regards,
Rimantas

Nice one, Hirofumi-shi!

Regards,
Jordan

On 20.09.2006 10:56, Rimantas L. wrote:

a.sort{ |i,j| [i.upcase, i] <=> [j.upcase, j] }

a.sort_by{ |i| [i.upcase, i] }

Great! I guess we have the winner

+1

``robert``

Hi,

WATANABE Hirofumi wrote:

Hi,

“Rimantas L.” [email protected] writes:

a.sort{ |i,j| [i.upcase, i] <=> [j.upcase, j] }

a.sort_by{ |i| [i.upcase, i] }

Wow!
This is what I was looking for.
Why couldn’t I think of this?

Thank you very much.

Sam