Indexing system - ruby newbie


#1

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!)


#2

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


#3

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"

#4

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”


#5

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”


#6

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


#7

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).


#8

how about that?

(n < 1) ? “_” : ((“A”[0] + n).chr)

n = 12

M


#9

another one:

tab = ["_", (“A”…“Z”).to_a].flatten
puts tab[n]


#10

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

#11

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


#12

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’ …


#13

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


#14

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 :
% 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


#15

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 :
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


#16

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


#17

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


#18

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


#19

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