Forum: Ruby Indexing system - ruby newbie

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
Adam G. (Guest)
on 2006-03-01 15:11
Hi there,

I'm trying to write a class which converts a number into letters like
so:
0  => -
1  => A
10 => J
27 => AA


... ad infinitum. My class looks like this at the moment (please don't
laugh!)

def letter(number)
  @index = number - 1
  if @index == 0
    return "-"
  end
  @index_string = ""
  @index_array= []
  while @index > 0 do
    @remainder = @index%27
    @index_array << @remainder
    @index = @index/27
  end
  @index_array
  @index_array.each do |i|
  # I'm sure there's a better way to do this
   @alphabet =
["A","B","C","D","E","F","G","H","I","J","K","L","M","N","O","P","Q","R","S","T","U","V","W","X","Y","Z"]
    @index_string << @alphabet[i-1]
  end
  return @index_string.reverse!
end

This works fine for the first round: 26 returns "Z". But 27 returns "AZ"
because @index_array is [0,1]. I'd appreciate any help (and tips on how
to write tighter code!)
Dave B. (Guest)
on 2006-03-01 15:34
(Received via mailing list)
Adam G. wrote:
> Hi there,
>
> I'm trying to write a class which converts a number into letters like
> so:
> 0  => -
> 1  => A
> 10 => J
> 27 => AA

Try this:

n = ""
while(n > 0)
  s << ?A + n % 26 - 1
  n /= 26
end
n << "-" if n.empty?
s.reverse

Also, your @alphabet is ("a".."z").to_a

Cheers,
Dave
Ross B. (Guest)
on 2006-03-01 16:00
(Received via mailing list)
On Wed, 2006-03-01 at 22:33 +0900, Dave B. wrote:
> Try this:
>
> n = ""
> while(n > 0)
>   s << ?A + n % 26 - 1
>   n /= 26
> end
> n << "-" if n.empty?
> s.reverse
>
> Also, your @alphabet is ("a".."z").to_a

This isn't as good as Dave's (it's potentially *lots* slower for a
start) but, well, I'm just an #inject addict really...:

	def letter(n)
	  (n < 1) ? '_' : (1...n).inject("A") { |curr, i| curr.succ }
	end

	letter(0)
	# => "_"
	letter(1)
	# => "A"
	letter(10)
	# => "J"
	letter(27)
	# => "AA"
	letter(397)
	# => "OG"
Adam G. (Guest)
on 2006-03-01 16:08
Hey thanks guys. I'm really enjoying learning ruby - especially because
the ruby community is so helpful.



Ross B. wrote:
> On Wed, 2006-03-01 at 22:33 +0900, Dave B. wrote:
>> Try this:
>>
>> n = ""
>> while(n > 0)
>>   s << ?A + n % 26 - 1
>>   n /= 26
>> end
>> n << "-" if n.empty?
>> s.reverse
>>
>> Also, your @alphabet is ("a".."z").to_a
>
> This isn't as good as Dave's (it's potentially *lots* slower for a
> start) but, well, I'm just an #inject addict really...:
>
> 	def letter(n)
> 	  (n < 1) ? '_' : (1...n).inject("A") { |curr, i| curr.succ }
> 	end
>
> 	letter(0)
> 	# => "_"
> 	letter(1)
> 	# => "A"
> 	letter(10)
> 	# => "J"
> 	letter(27)
> 	# => "AA"
> 	letter(397)
> 	# => "OG"
Adam G. (Guest)
on 2006-03-01 17:14
Dear Ross,

it works a treat but I'm having a bit of trouble figuring out what's
going on.

 	  (n < 1) ? '_' : (1...n).inject("A") { |curr, i| curr.succ }

I get this:
if n<1
  '_'
else

But I'm stuck here.

  (1...n).inject("A") { |curr, i| curr.succ}

I still can't quite get my head around blocks beyond .each do |x|


Ross B. wrote:
> On Wed, 2006-03-01 at 22:33 +0900, Dave B. wrote:
>> Try this:
>>
>> n = ""
>> while(n > 0)
>>   s << ?A + n % 26 - 1
>>   n /= 26
>> end
>> n << "-" if n.empty?
>> s.reverse
>>
>> Also, your @alphabet is ("a".."z").to_a
>
> This isn't as good as Dave's (it's potentially *lots* slower for a
> start) but, well, I'm just an #inject addict really...:
>
> 	def letter(n)
> 	  (n < 1) ? '_' : (1...n).inject("A") { |curr, i| curr.succ }
> 	end
>
> 	letter(0)
> 	# => "_"
> 	letter(1)
> 	# => "A"
> 	letter(10)
> 	# => "J"
> 	letter(27)
> 	# => "AA"
> 	letter(397)
> 	# => "OG"
Edward F. (Guest)
on 2006-03-01 17:50
(Received via mailing list)
On Thu, Mar 02, 2006 at 12:14:47AM +0900, Adam G. wrote:
> But I'm stuck here.
>
>   (1...n).inject("A") { |curr, i| curr.succ}
>
> I still can't quite get my head around blocks beyond .each do |x|

#inject is one of the basic tools of functional programming.  That's
why it seems hard.  It's a very different way of thinking.  And also
very powerful.

That said, I think this case is a completely gratuitious use of
inject.  The tipoff is that the argument "i" is completely ignored.

I like this much better:

def letter(n)
  return '_' if n==0
  n==1 ? "A" : letter(n-1).succ
end

regards,
Ed
Ross B. (Guest)
on 2006-03-01 18:27
(Received via mailing list)
On Thu, 2006-03-02 at 00:50 +0900, Edward F. wrote:
>
> That said, I think this case is a completely gratuitious use of
> inject.  The tipoff is that the argument "i" is completely ignored.
>

Well, sorry, I didn't realise we had to use them all. I like to use it
where I want to give back something new from a block, but don't want:

	a = []
	something.each { |e| a << e end }
	a

In this case it was just a snazzier alternative to doing the (n-1).times
and so on...

Did I say I'm an #inject *addict* ?

> I like this much better:
>
> def letter(n)
>   return '_' if n==0
>   n==1 ? "A" : letter(n-1).succ
> end
>

Well, to each his own, but (maybe this is a bit pathological,
though...):

	def letter(n)
	  (1...n).inject("A") { |curr, i| curr.succ}
	end

	def letter2(n)
	  return '_' if n==0
	  n==1 ? "A" : letter2(n-1).succ
	end

	p letter(327021)
	# => "ROSS"

	p letter2(327021)
	# => -:3:in `letter2': stack level too deep (SystemStackError)
                    from -:3:in `letter2'
	            from -:11
Ross B. (Guest)
on 2006-03-01 18:37
(Received via mailing list)
On Thu, 2006-03-02 at 00:14 +0900, Adam G. wrote:
> else
>
> But I'm stuck here.
>
>   (1...n).inject("A") { |curr, i| curr.succ}
>
> I still can't quite get my head around blocks beyond .each do |x|

Inject is real easy, and very handy. It's just like 'each', except it
also allows the result of the previous iteration to be injected via the
first argument. For the first iteration, you provide the initial result.

For example:

	a = [1,2,3,4,5]

	a.inject(0) { |sum, i| sum + i }
	# => 15

What happens is:

	Block is called with sum = 0, i = 1
	  Block returns 1
	Block is called with sum = 1, i = 2
	  Block returns 3
	Block is called with sum = 3, i = 3
	  Block returns 6
	Block is called with sum = 6, i = 4
	  Block returns 10
	Block is called with sum = 10, i = 5
	  Block returns 15
	No more elements, so inject returns 15.

In Ruby, inject allows you to omit the initial value, in which case the
first _two_ elements from the enumerable are passed to the first
iteration, with things proceeding as above from there, so I could have
written:

	a.inject { |sum, i| sum + i }

And would have:

	Block is called with sum = 1, i = 2
	  Block returns 3
	Block is called with sum = 3, i = 3
	  Block returns 6
	.
	.
	etc.

You can use inject for much more than just summing stuff up. Comes in
very handy for these cryptic one-lines (even if you have to 'misuse' it
a bit occasionally).
Peter E. (Guest)
on 2006-03-01 18:43
(Received via mailing list)
how about that?

(n < 1) ? "_" : (("A"[0] + n).chr)

n = 12
> M
Peter E. (Guest)
on 2006-03-01 18:46
(Received via mailing list)
another one:

tab = ["_", ("A".."Z").to_a].flatten
puts tab[n]
James G. (Guest)
on 2006-03-01 18:49
(Received via mailing list)
On Mar 1, 2006, at 10:42 AM, Peter E. wrote:

> how about that?
>
> (n < 1) ? "_" : (("A"[0] + n).chr)

And we can shorten "A"[0] to ?A.

James Edward G. II
Ross B. (Guest)
on 2006-03-01 18:50
(Received via mailing list)
On Thu, 2006-03-02 at 01:42 +0900, Peter E. wrote:
> how about that?
>
> (n < 1) ? "_" : (("A"[0] + n).chr)
>
> n = 12
> > M
>

n = 27
> \

OP wanted:

 ...  24   25   26   27    28    29   ...
 ... 'X', 'Y', 'Z', 'AA', 'AB', 'AC' ...
Adam S. (Guest)
on 2006-03-01 19:43
(Received via mailing list)
> OP wanted:
>
>  ...  24   25   26   27    28    29   ...
>  ... 'X', 'Y', 'Z', 'AA', 'AB', 'AC' ...

Here's one that uses succ to do the dirty work, but doesn't complicate
things with inject:
(note that '@'.succ = 'A')

def letter n
  l='@'
  n.times{l.succ!}
  l.gsub(/@/,'-')
end

-Adam
Logan C. (Guest)
on 2006-03-01 23:43
(Received via mailing list)
On Mar 1, 2006, at 12:43 PM, Adam S. wrote:

>   l='@'
>   n.times{l.succ!}
>   l.gsub(/@/,'-')
> end
>
> -Adam
>

And now for the over-engineered approach to balance out the golfing <g>:
% cat indexer.rb
class Indexer
   def initialize
     @index_cache = ('A'..'Z').to_a
     @index_cache.unshift('-')
   end

   def alpha_index(i)
     if res = @index_cache[i]
       res
     else
       @index_cache[i] = alpha_index(i - 1).succ
     end
   end
   alias [] alpha_index
end

if $0  == __FILE__
   idx = Indexer.new
   puts idx.alpha_index(27)
   puts idx.alpha_index(0)
   puts idx[26]
end

% ruby indexer.rb
AA
-
Z
William J. (Guest)
on 2006-03-02 04:28
(Received via mailing list)
Adam G. wrote:
> .. ad infinitum. My class looks like this at the moment (please don't
>     @remainder = @index%27
>   return @index_string.reverse!
> end
>
> This works fine for the first round: 26 returns "Z". But 27 returns "AZ"
> because @index_array is [0,1]. I'd appreciate any help (and tips on how
> to write tighter code!)
>
>
> --
> Posted via http://www.ruby-forum.com/.


def letter( n )
  return "-"   if n == 0

  result = n.to_s(27)
  i = 0
  while i < result.size - 1
    inc = result[-i-2,1].to_i(27)
    inc += 1   if inc + result[-1,1].to_i(27) > 26
    n += 27**i * inc
    result = n.to_s(27)
    i += 1
  end

  result.tr(((1..9).to_a + ('a'..'q').to_a).join,
             ('A'..'Z').to_a.join )
end
A LeDonne (Guest)
on 2006-03-02 22:45
(Received via mailing list)
On 3/1/06, Logan C. <removed_email_address@domain.invalid> wrote:
> > (note that '@'.succ = 'A')
> And now for the over-engineered approach to balance out the golfing <g>:
>      else
>    puts idx[26]
> end
>
> % ruby indexer.rb
> AA
> -
> Z
>
>

A different approach from others in this thread... no memoization, no
math in the method (let to_s(base) handle it), and a loop that only
runs, at most, as many times as the length of the resulting string.
The idea is, change to base 26, then "uncarry" the ones. Should be
fast.


def letterize(num)
    b26 = num.to_s(26).tr("0-9a-p","@-Y")
    while b26.sub!(/.@/) { |s| (s[0]-1).chr + "Z" }
    end
    b26.sub!(/^@/, "")
    b26 << "-" if b26.empty?
    b26
end
Logan C. (Guest)
on 2006-03-03 00:26
(Received via mailing list)
On Mar 2, 2006, at 3:43 PM, A LeDonne wrote:

>>> complicate
>>>
>>    def alpha_index(i)
>>    idx = Indexer.new
>>
>     while b26.sub!(/.@/) { |s| (s[0]-1).chr + "Z" }
>     end
>     b26.sub!(/^@/, "")
>     b26 << "-" if b26.empty?
>     b26
> end
>

Thank you, I was trying hard to work out a way to use to_s(26) and
couldn't quit get it to work
Malte M. (Guest)
on 2006-03-03 17:09
(Received via mailing list)
Adam G. schrieb:
> I'm trying to write a class which converts a number into letters like
> so:
> 0  => -
> 1  => A
> 10 => J
> 27 => AA

Only after reading the other posts, I realized you want Y, Z, AA, AB, AC
and not Y, Z, AA, BB, CC, which is what is produced by this:

def letter(i)
  return '-' if i < 1
  ('A'..'Z').to_a.at(i%26 - 1) * ((i-1)/26 + 1)
end

Malte
Rob (Guest)
on 2006-03-08 04:15
(Received via mailing list)
On 2006-03-03, Malte M. <removed_email_address@domain.invalid> wrote:
> Only after reading the other posts, I realized you want Y, Z, AA, AB, AC
> and not Y, Z, AA, BB, CC, which is what is produced by this:
>
> def letter(i)
>   return '-' if i < 1
>   ('A'..'Z').to_a.at(i%26 - 1) * ((i-1)/26 + 1)
> end
>
> Malte

could add one method  to Integer
and avoid a separate class.
(_to_aa is a helper that works on
0..25 nstead of 1..26)

eg
puts 10.to_aa, (10**1000).to_aa

class Integer

	def to_aa # (1..26) to (A..Z)
	  if (self < 1)
	    	'-'
    else
      (self-1)._to_aa
	  end
	end

  protected
  def _to_aa # helper (0..25) to (A..Z)
	  if  (self < 26)
	  	(self+?A).chr
	  else
		  ((self/26)-1)._to_aa+(self%26)._to_aa
	  end
  end

end
This topic is locked and can not be replied to.