The Elixir Caesar Cipher


Write an elixir Caesar Cipher program that will encrypt a message. Your program should prompt the user for input (IO.gets), and they will enter the following:

number,and a message
  • number - This is the number of characters to shift the Caesar Cipher e.g. 3 will shift the alphabet to xyzabcdefghijklmnopqrstuvw
  • message - This is alphabetic, lowercase and whitespace is permitted (whitespace will simply map to whitespace of the encoded text).

The output will be lowercase with whitespace.

For example:

iex> --enter your code here--
3,abc defg hijklmnopqrstuvwxyz
xyz abcd efghijklmnopqrstuvw

The solution that completes this puzzle with the shortest number of characters wins.


So this weeks elixirgolf probably needed a better definition as there were some queries around cycling of the letters, and the bounds of the numbers passed in etc. However it did lead to some very creative answers and as our goal is to use these puzzles as a learning tool, we are going to be lenient.


There was a lot of buzz around the Caesar cipher puzzle. Many of the solutions were very creative and influenced one another. As a result the winning solution was an evolution of ideas. Therefore we will give credit to all those who contributed.


All though not initially the winner @henrik persevered and managed to produce a working Caesar cipher in a staggering 103 characters! His solution was built from his own concepts and inspiration from both @gregvaughn and @MPAhrens.

Winner: @henrik

char count: 103

[n,p]=IO.gets("")|>String.split",";IO.puts for<<c<-p>>,do: c<97&&c||97+rem c-71-String.to_integer(n),26

@henrik uses the same approach as @gregvaughn to extract the shift number and message, as well as calculating the mapping cycle (see Greg’s solution for a detailed description).

...for<<c<-p>>,do:... Henrik uses the trick @MPAhrens developed to convert the message p, into a character list. He does this by wrapping the entire comprehension body in a binaray e.g. << "a" >>. Notice how the comprehension body c <- p is wrapped in the binary, this saves 12 characters.

Well done and also a big well done to both @gregvaughn and @MPAhrens

Second Place

@gregvaughn was in the pole position with his solution. Although the rules were a bit fuzzy around the bounds of the “shift” number, his solution met the basic criteria. It also influenced the winning solution.

Second: @gregvaughn

char count: 114

# for an offset range of -26 to + 26
[n,p]=IO.gets("")|>String.split",";IO.puts for c<-to_char_list(p),do: c<97&&c||97+rem c-71-String.to_integer(n),26

First this solution gets the user’s input IO.gets("")|>String.split",", which will be the String 3,abc def and uses the |> operator to pass this String value to the String.split/2 function. Once split around the , the number is bound to n and the message is bound to the p variable ([n,p]).

The algorithm then takes the message p and converts it into a character list, which is just a list of ASCII integers to_char_list(p). So now @gregvaughn uses comprehension for c<-to_char_list(p),do:... to iterate over each of the integer values binding them to the variable c.

The real work of the algorithm takes place in the body of the comprehension, where @gregvaughn calculates the new value of the ASCII integer based on the “shift” number c<97&&c||97+rem c-71-String.to_integer(n),26. First the algorithm checks to see if the ASCII value is less than 97 (the ASCII number for the letter a) if it is then it must be a space or other none alphabetic character, if it is then that character is returned c<97&&c. The logic reads: if c is less than 97 then return c.

If c is great than 97 then the algorithm must remap the ASCII character value based on the “shift” number n, this is triggered from the || (or) operator. To remap the original letter to the new letter we need to increment the ASCII value by our value of n, but we need to handle values great than 26 (the number of letters in the Alphabet).

First we take the base value of a which is 97 and add our “shift” value by converting our ASCII letter value to a value in the range of 26 using c-71, we then subtract our “shift” number from this value while converting it to an Integer String.to_integer(n). Finally we find the remainder of 26 using rem c-71-String.to_integer(n),26, and add it to 97. At last we have the mapped value of c and we convert this character list to a String using IO.puts. If the value of n is out of the range -26..+26 then we get other printable characters e.g. when n=30 output is ]^_ `ab

A great solution thanks.

Honourable mentions

Submitted by @andrei_here:

char count: 88

{_,[n,c]}=:io.fread"",'~d,~s';:io.format"~s",[(for x<-c,do: x>96&&rem(x-97+n,26)+97||x)]

This solution isn’t a winner but is definately in the spirit of elixirgolf, which is to use puzzles for learning the elixir language.

Unfortunately this solution doesn’t cycle the correct way i.e. abc should map to xyz, when the shift number is 3. Also the solution can only handle single words i.e. no space separation. However it does have some very interesting concepts:

{_,[n,c]}=:io.fread"",'~d,~s'; Uses the Erlang io.fread function that reads characters from the standard input. The format of the input is defined as '~d,~s' which defines digits, a comma and a single string. The result of this input is bound to the List variable n (for digits) and c (for the message) the :ok atom which io.fread returns is ignored with the wildcard _. Note that the value of c is a character list e.g. 'abc'.

@andrei_here then uses the character list to change the mapping of the characters. As a character list is just a list of integers he is able to use a comprehension to get each value and “shift” it by 26 characters (for x<-c,do: x>96&&rem(x-97+n,26)+97||x). As we still have our data as a character list @andrei_here uses the Erlang function :io.format to format the character list into a string, note that it is wrapped in square brackets.

Submitted by @gregvaughn

char count: 114

# for an offset range of -26 to + 26
# 114 characters
[n,p]=IO.gets("")|>String.split",";IO.puts for c<-to_char_list(p),do: c<97&&c||97+rem c-71-String.to_integer(n),26

# for any offset range
# 122 characters
[n,p]=IO.gets("")|>String.split",";IO.puts for c<-to_char_list(p),do: c<97&&c||97+rem c-71-rem(String.to_integer(n),26),26

Submitted by @henrik

char count: 124

[n,s]=String.split IO.gets(""),",";IO.write Regex.replace~r/\S/,s,fn<<c>>->q=c-rem String.to_integer(n),26;q<?a&&26+q||q;end
# If we combine "my" Regex.replace with Greg Vaughn's smarter maths, we get to
# 121 together:
[n,s]=String.split IO.gets(""),",";IO.write Regex.replace~r/\S/,s,fn<<c>>->?a+rem c-71-rem(String.to_integer(n),26),26end

# Just for fun: Using Stream.cycle as suggested by Martin Svalin:
[n,s]=String.split IO.gets(""),",";IO.write Regex.replace~r/\S/,s,fn<<c>>-> Stream.cycle(?z..?a),?z-c+String.to_integer(n)end

# Just for fun: Using Code.eval_string to convert to int and char list at the
# same time… (The "do" block contents are borrowed from Greg Vaughn)
{s,n: n}=Code.eval_string String.replace(IO.gets(""),~r/(.+),(.*)\n/,"n=\\1;'\\2'");IO.puts for c<-s,do: c<97&&c||97+rem c-71-rem(n,26),26

# Just for fun: if we limit to single digits we can do:
<<n,?,,s::bits>>=IO.gets"";IO.write Regex.replace~r/\S/,s,fn<<c>>->97+rem c-71-n+?0,26end

# or even
<<n,?,,s::bits>>=IO.gets"";IO.write Regex.replace~r/\S/,s,fn<<c>>->97+rem c-n-23,26end

# If you don't need to "wrap" (so you can do d->a but not a->x).
# `import String; … split … to_integer` is the exact same char count as not
# importing.
import String;[n,s]=split IO.gets(""),",";IO.write Regex.replace~r/\S/,s,fn<<c>>->c-to_integer(n)end

For more details see @henrik’s gist:

Submitted by @MPAhrens

char count: 131

fn[n,m]->for<<x<-m>>,into: "",do: <<(if x<?a,do: 32,else: rem(x+String.to_integer(n)-?a,26)+?a)>>end.(String.split IO.gets(""),",")

The solution provided by @MPAhrens works as follows. He creates a giant anonymous function that he calls, inline, with an IO.gets "" call which he splits. The function broken down looks like this:

fn [n,m] -> [n,m] end.(String.split IO.gets(""),"," )

He then uses a comprehension that uses the very interesting technique of wrapping the entire body in a binary <<x<-m>> this has the benfit of converting the message string m to a character list and binding each integer to x. His comprehension also uses the into: "" function, which returns a function that collects values alongside the initial accumulation value. Basically for each iteration of the comprehension the character value is appended to the string "".

The comprehension body then does the mapping calculation:

<<(if x<?a,do: 32,else: rem(x+String.to_integer(n)-?a,26)+?a)>>

Again he wraps the entire body in a binary <<>>, then he uses an if/else function to determine if the message character value is below the letter a (ASCII 97) if it is, it assumes that it is a space character and the ASCII value 32 is returned. Otherwise @MPAhrens calculates the new character mapping.

The calculation finds the remainder rem(X_VALUE, 26) of the current character ASCII value x and adds the “shift” number (while converting it to an integer) less the ASCII value of ?a. The result is then added to the ASCII value of ?a.

Great solution and we really liked the interesting use of binaries.


comments powered by Disqus