I want to encode an id to unique strings containing numbers and uppercase letters, like this:
40 => A5TY8
but
41 => Y7HEG
Where the output for 41 is completely different from 40's A5TY9.
How to do this?
Asked
Active
Viewed 2,287 times
2
the Tin Man
- 158,662
- 42
- 215
- 303
mitch
- 2,235
- 3
- 27
- 46
-
How many ID numbers are you anticipating? Is it feasible to make a list just once, or do you need to be able to generate on the fly? – Floris Jan 21 '13 at 18:18
-
I want generate it on the fly. I want 2 functions encode and decode, so it can't be md5 or sha. It can be billions of ids so generating once and using it - it's not the solution. – mitch Jan 21 '13 at 18:23
-
Do they have to be unique and short? If so you need some kind of look-up table. – tadman Jan 21 '13 at 18:31
-
Not necessarily short. Unique, decodeable – mitch Jan 21 '13 at 18:32
-
http://uuidtools.rubyforge.org/ – ChuckE Jan 21 '13 at 18:43
-
1Sounds similar to this problem http://stackoverflow.com/questions/742013/how-to-code-a-url-shortener – Pafjo Jan 21 '13 at 18:56
-
I tried this before but outputs are similar. JRE, JRF... – mitch Jan 21 '13 at 18:58
-
@MichałK. Try to randomize the alphabet and test again. – Pafjo Jan 21 '13 at 19:24
-
1@Pafjo I did it. And I have: TA => 494 T0 => 495 TD => 496 TW => 497 T3 => 498 T9 => 499 – mitch Jan 21 '13 at 19:31
1 Answers
3
I wanted to tell you all about digests.. but seeing your comment I think you want something like this
https://github.com/namick/obfuscate_id
or better yet a more generic
https://github.com/patdeegan/integer-obfuscator
you can also encode it with Base64 which is reverisble
http://ruby-doc.org/stdlib-1.9.3/libdoc/base64/rdoc/Base64.html
note that you might want to use urlsafe_encode64 so you won't have /n and stuff in there
so you can do something like
require "base64"
original = 41
converted = Base64.urlsafe_encode64("41")
converted_for_display = converted.tr('^A-Za-z0-9', '')
# => "NDE"
reversed = Base64.urlsafe_decode64(converted)
you can also encrypt the number which should make it unique and reversible, but slower and with more hassle but you can pick whatever key you want
Nick Ginanto
- 31,090
- 47
- 134
- 244
-
Thank you, but solution numbers with base are similar, NDE, NDA. The most interesting is using integer-obfuscator and then use base. – mitch Jan 21 '13 at 18:50
-
ALPHABET = "ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789".split("").shuffle.join def bijective_encode(i) return ALPHABET[0] if i == 0 s = '' base = ALPHABET.length while i > 0 s << ALPHABET[i.modulo(base)] i /= base end s.reverse end def bijective_decode(s) i = 0 base = ALPHABET.length s.each_char { |c| i = i * base + ALPHABET.index(c) } i end $salt = 1369136 def decode(id) (bijective_decode(id) >> 18) - $salt end def encode(id) bijective_encode((id+$salt) << 18) end I can't "Answer your question" but this is the best solution for me. – mitch Jan 24 '13 at 10:47