Tips for golfing in Ruby

Question

What general tips can you give for golfing in Ruby?

I'm looking for ideas that can be applied to code golf problems in general that are specific to Ruby. (For example, "Remove comments" would not be an answer.)

Please post one tip per answer.

Answer 1

• The numbers 100 to 126 can be written as ?d to ?~ in 1.8.
• On a similar note if you need a single-character string in 1.9 ?x is shorter than "x".
• If you need to print a string without appending a newline, $><<"string" is shorter than print"string".
• If you need to read multiple lines of input $<.map{|l|...} is shorter than while l=gets;...;end. Also you can use $<.read to read it all at once.
• If you're supposed to read from a file, $< and gets will read from a file instead of stdin if the filename is in ARGV. So the golfiest way to reimplement cat would be: $><<$<.read.

Answer 2

Use the splat operator to get the tail and head of an array:

head, *tail = [1,2,3]
head => 1
tail => [2,3]

This also works the other way:

*head, tail = [1,2,3]
head => [1,2]
tail => 3

Use the * method with a string on an array to join elements:

[1,2,3]*?,
=> "1,2,3"

Answer 3

• Use abort to terminate the program and print a string to STDERR - shorter than puts followed by exit
• If you read a line with gets, you can then use ~/$/ to find its length (this doesn't count a trailing newline if it exists)
• Use [] to check if a string contains another: 'foo'['f'] #=> 'f'
• Use tr instead of gsub for character-wise substitutions: '01011'.tr('01','AB') #=> 'ABABB'
• If you need to remove trailing newlines, use chop instead of chomp

Answer 4

End your end.

Try to remove end from your code.

Don't use def...end to define functions. Make a lambda with the new -> operator in Ruby 1.9. (The -> operator is a "stabby lambda", or "dash rocket".) This saves 5 characters per function.

# 28 characters
def c n
/(\d)\1/=~n.to_s
end

# 23 characters, saves 5
c=->n{/(\d)\1/=~n.to_s}

Method calls are c n or c(n). Lambda calls are c[n]. Changing each c n to c[n] costs 1 character, so if you can use c n more than 5 times, then keep the method.

All methods that take do...end blocks can take {...} blocks instead. This saves 3 to 5 characters. If the precedence of {...} is too high, then use parentheses to fix it.

# 48 characters
(?a..?m).zip (1..5).cycle do|a|puts a.join','end

# WRONG: passes block to cycle, not zip
(?a..?m).zip (1..5).cycle{|a|puts a.join','}

# 45 characters, saves 3
(?a..?m).zip((1..5).cycle){|a|puts a.join','}

Replace if...else...end with the ternary operator ?:. If a branch has two or more statements, wrap them in parentheses.

# 67 characters
if a<b
puts'statement 1'
puts'statement 2'else
puts'statement 3'end

# 62 characters, saves 5
a<b ?(puts'statement 1'
puts'statement 2'):(puts'statement 3')

You probably don't have while or until loops, but if you do, then write them in modifier form.

(a+=1
b-=1)while a<b

Answer 5

Addition to w0lf

When working with arrays, .compact can be replaced with -[nil] to save 2 chars.

Combined with above -> you can make it even shorter with -[p] to save another 2 chars.

Answer 6

Use the short predefined variables wherever possible, e.g. $* instead of ARGV. There's a good list of them here, along with a lot of other useful information.

Answer 7

Use operator methods instead of parentheses

Let's say you want to express a*(b+c). Because of precedence, a*b+c won't work (obviously). Ruby's cool way of having operators as methods comes to the rescue! You can use a.*b+c to make the precedence of * lower than that of +.

a*(b+c) # too long
a*b+c   # wrong
a.*b+c  # 1 byte saved!

This can also work with the ! and ~ operators (things like unary + or unary - don't work because their methods are -@ and +@, saving () but adding .@)

(~x).to_s # too long
~x.to_s   # error
x.~.to_s  # 1 byte saved!

Answer 8

Don't use the true and false keywords.

Use:

• !p for true (thanks, histocrat!)
• !0 for false. If all you need is a falsy value, then you can simply use p (which returns nil).

to save some chars.

Answer 9

When you are using string interpolation, (as you should pr Martin Büttner's post), you don't need the curly brackets if your object has a sigil ($, @) in front of it. Useful for magical variables like $_, $&, $1 etc:

puts "this program has read #$. lines of input"

So also if you need to print a variable more than you use it otherwise, you may save some bytes.

a=42; puts "here is a: #{a}"; puts "here is a again: #{a}"
$b=43; puts "here is b: #$b"; puts "here is b again: #$b"

Answer 10

New features in Ruby 2.3 and 2.4

It's good to stay abreast of new language features that will help your golf game. There are a few great ones in the latest Rubies.

Ruby 2.3

The safe navigation operator: &.

When you call a method that might return nil but you want to chain additional method calls if it's not, you waste bytes handling the nil case:

arr = ["zero", "one", "two"]
x = arr[5].size
# => NoMethodError: undefined method `size' for nil:NilClass

x = arr[5].size rescue 0
# => 0

The "safe navigation operator" stops the chain of method calls if one returns nil and returns nil for the whole expression:

x = arr[5]&.size || 0
# => 0

Array#dig & Hash#dig

Deep access to nested elements, with a nice short name:

o = { foo: [{ bar: ["baz", "qux"] }] }
o.dig(:foo, 0, :bar, 1) # => "qux"

Returns nil if it hits a dead end:

o.dig(:foo, 99, :bar, 1) # => nil

Enumerable#grep_v

The inverse of Enumerable#grep—returns all elements that don't match the given argument (compared with ===). Like grep, if a block is given its result is returned instead.

(1..10).grep_v 2..5 # => [1, 6, 7, 8, 9, 10]
(1..10).grep_v(2..5){|v|v*2} # => [2, 12, 14, 16, 18, 20]

Hash#to_proc

Returns a Proc that yields the value for the given key, which can be pretty handy:

h = { N: 0, E: 1, S: 2, W: 3 }
%i[N N E S E S W].map(&h)
# => [0, 0, 1, 2, 1, 2, 3]

Ruby 2.4

Ruby 2.4 isn't out yet, but it will be soon and has some great little features. (When it's released I'll update this post with some links to the docs.) I learned about most of these in this great blog post.

Enumerable#sum

No more arr.reduce(:+). You can now just do arr.sum. It takes an optional initial value argument, which defaults to 0 for Numeric elements ([].sum == 0). For other types you'll need to provide an initial value. It also accepts a block that will be applied to each element before addition:

[[1, 10], [2, 20], [3, 30]].sum {|a,b| a + b }
# => 66

Integer#digits

This returns an array of a number's digits in least-to-greatest significance order:

123.digits # => [3, 2, 1]

Compared to, say, 123.to_s.chars.map(&:to_i).reverse, this is pretty nice.

As a bonus, it takes an optional radix argument:

a = 0x7b.digits(16) # => [11, 7]
a.map{|d|"%x"%d} # => ["b", "7"]

Comparable#clamp

Does what it says on the tin:

v = 15
v.clamp(10, 20) # => 15
v.clamp(0, 10) # => 10
v.clamp(20, 30) # => 20

Since it's in Comparable you can use it with any class that includes Comparable, e.g.:

?~.clamp(?A, ?Z) # => "Z"

String#unpack1

A 2-byte savings over .unpack(...)[0]:

"👻💩".unpack(?U)    # => [128123]
"👻💩".unpack(?U)[0] # => 128123
"👻💩".unpack1(?U)   # => 128123

Precision argument for Numeric#ceil, floor, and truncate

Math::E.ceil(1) # => 2.8
Math::E.floor(1) # => 2.7
(-Math::E).truncate(1) # => -2.7

Multiple assignment in conditionals

This raises an error in earlier versions of Ruby, but is allowed in 2.4.

(a,b=1,2) ? "yes" : "no" # => "yes"
(a,b=nil) ? "yes" : "no" # => "no"

Answer 11

If you need to find if a particular element e is inside a range r, you can use

r===e

instead of the longer:

r.cover?(e) # only works if `r.exclude_end?` is false

or

r.member?(e)

or

r.include?(e)

Answer 12

Build arrays using a=i,*a to get them in reverse order. You don't even need to initialize a, and if you do it doesn't have to be an array.

Answer 13

Avoid length in if a.length<n

length is 6 bytes, a bit costly in code golf. in many situations, you can instead check if the array has anything at a given point. if you grab past the last index you will get nil, a falsey value.

So you can Change:

if a.length<5 to if !a[4] for -5 bytes

or

if a.length>5 to if a[5] for -6 bytes

or

if a.length<n to if !a[n-1] for -3 bytes

or

if a.length>n to if a[n] for -6 bytes

Note: will only work with an array of all truthy values. having nil or false within the array may cause problems.

Answer 14

New features in Ruby 2.7 (experimental)

Ruby 2.7 is in prerelease (as of 17 Jun 2019) and has some features that look great for golfing. Note that some of them might not make it into the final 2.7 release.

All changes in Ruby 2.7-preview1: https://github.com/ruby/ruby/blob/v2_7_0_preview1/NEWS

Numbered block parameters

This is my favorite. It lets you finally drop the |a,b| in a block:

Edit: The syntax was changed from @1 to _1.

%w[a b c].zip(1..) { puts _1 * _2 }
# => a
#    bb
#    ccc

Method reference operator: .:

Edit: This was unfortunately removed before release.

.: is syntactic sugar for the .method method, e.g.:

(1..5).map(&1r.:/)
# => [(1/1), (1/2), (1/3), (1/4), (1/5)]

Pattern matching

I'm not sure how much use this will see in golf, but it's a great feature for which I only have a contrived example:

def div(*a)
  case a
    in [0, 0] then nil
    in [x, 0] if x > 0 then Float::INFINITY
    in [x, 0] then -Float::INFINITY
    in [x, y] then x.fdiv(y)
  end
end

div(-3, 0) # => -Infinity

The pattern matching syntax has lots of features. For a complete list, check out this presentation: https://speakerdeck.com/k_tsj/pattern-matching-new-feature-in-ruby-2-dot-7

This is also the feature most likely to change before 2.7 is finished; it even prints a warning when you try to use it, which you should heed:

warning: Pattern matching is experimental, and the behavior may change in future versions of Ruby!

Beginless Range: ..3

Analogous to the endless Range introduced in 2.6, it may or may not have much use in golfing:

%w[a v o c a d o].grep(..?m)
# => ["a", "c", "a", "d"]

Enumerable#tally to count like elements

This could be useful in golfing:

%w[a v o c a d o].tally
# => {"a"=>2, "v"=>1, "o"=>2, "c"=>1, "d"=>1}

Enumerable#filter_map to filter+map in one

(1..20).filter_map {|i| 10 * i if i.even? }
# => [20, 40, 60, 80, 100]

If the block returns nil or false the element will be omitted from the result.

Integer#[] takes a second argument or range:

You've long been able to get a specific bit from an integer with with subscript notation:

n = 77 # (binary 01001101)
n[3] # => 1

Now you can get the value of a range of bits by a second length argument or a range.

n = 0b01001101
n[2, 4] # => 3 (0011)
n[2..5] # => 3

Note that bits are indexed from least- to most-significant (right to left).

Answer 15

$_ is last read line.

• print - if no argument given print content of $_
• ~/regexp/ - short for $_=~/regexp/

In Ruby 1.8, you have four methods in Kernel that operate on $_:

• chop
• chomp
• sub
• gsub

In Ruby 1.9, these four methods exist only if your script uses -n or -p.

If you want to print some variable often then use trace_var(:var_name){|a|p a}

Answer 16

Use string interpolation!

1. To replace to_s. If you need parentheses around whatever you want to turn into a string, to_s is two bytes longer than string interpolation:

   (n+10**i).to_s
   "#{n+10**i}"
   

2. To replace concatenation. If you concatenate something surrounded by two other strings, interpolation can save you one byte:

   "foo"+c+"bar"
   "foo#{c}bar"
   

   Also works if the middle thing is itself concatenated, if you just move the concatenation inside the interpolation (instead of using multiple interpolations):

   "foo"+c+d+e+"bar"
   "foo#{c+d+e}bar"
   

Answer 17

Don't use #each. You can loop over all elements just fine with #map. So instead of

ARGV.each{|x|puts x}

you can do the same in less bytes.

ARGV.map{|x|puts x}

Of course, in this case puts $* would be even shorter.

---

There are literals for rational and complex numbers:

puts 3/11r == Rational(3,11)
puts 3.3r == Rational(66,20)
puts 1-1.i == Complex(1,-1)

=> true
true
true

---

You can use most bytes within strings. "\x01" (6 bytes) can be shortened to "" (3 bytes). If you only need this one byte, this can be shortened even further to ? (2 bytes).

By the same token, you can get newlines shorter like this:

(0..10).to_a.join'
'

 => "0\n1\n2\n3\n4\n5\n6\n7\n8\n9\n10"

You can use ?\n and ?\t as well, which is one byte shorter than "\n" and "\t". For obfuscation, there also ?\s, a space.

---

Use constants instead of passing arguments around, even if you need to change them. The interpreter will give warnings to stderr, but who cares. If you need to define more variables related to each other, you can chain them like this:

A=C+B=7+C=9

=> A=17, B=16, C=9

This is shorter than C=9;B=16;A=17 or C=0;B=C+7;A=C+B.

---

If you need an infinite loop, use loop{...}. Loops of unknown length may be shorter with other loops:

loop{break if'
'==f(gets)}

while'
'!=f(gets);end

---

Some more gsub/regexp tricks. Use the special '\1' escape characters instead of a block:

"golf=great short=awesome".gsub(/(\w+)=(\w+)/,'(\1~>\2)')

"golf=great short=awesome".gsub(/(\w+)=(\w+)/){"(#{$1}~>#{$2})")

And the special variables $1 etc. if you need to perform operations. Keep in mind they are defined not only inside the block:

"A code-golf challenge." =~ /(\w+)-(\w+)/
p [$1,$2,$`,$']

=> ["code", "golf", "A ", " challenge."] 

---

Get rid of spaces, newlines, and parentheses. You can omit quite a bit in ruby. If in doubt, always try if it works without, and keep in mind this might break some editor syntax highlighting...

x+=1if$*<<A==????::??==??

Answer 18

If you ever need to get a number from ARGV, get, or something similar to do something that many times, instead of calling to_i on it, you can just use ?1.upto x{do something x times} where x is a string.

So using ?1.upto(a){} instead of x.to_i.times{} will save you 2 characters.

You can also re-write things like p 1 while 1 or p 1 if 1 as p 1while 1 or p 1if 1

That example isn't very useful, but it could be used for other things.

Also, if you need to assign the first element of an array to a variable, a,=c will save two characters as opposed to a=c[0]

Answer 19

Save some bytes when removing repeated elements of an array

a.uniq # before
a|[]   # after
    ^^

If you will be using an empty array [] in a variable, you can save even more bytes:

a.uniq;b=[] # before
a|b=[]      # after
      ^^^^^

Answer 20

Scientific notation can often be used to shave off a char or two:

x=1000
#versus
x=1e3

Answer 21

Kernel#p is a fun method.

Use p var instead of puts var. This works perfectly with integers and floats, but not with all types. It prints quotation marks around strings, which is probably not what you want.

Used with a single argument, p returns the argument after printing it.

Used with multiple arguments, p returns the arguments in an array.

Use p (with no arguments) instead of nil.

Answer 22

Yet another way to use the splat operator: if you want to assign a single array literal, a * on the left-hand side is shorter than brackets on the right-hand side:

a=[0]
*a=0

With multiple values you don't even need the splat operator (thanks to histocrat for correcting me on that):

a=[1,2]
a=1,2

Answer 23

To join an array, instead of this

[...].join

do this

[...]*''

which saves 2 bytes. To join with a separator use

[...]*?,

Answer 24

Save a byte when printing a word with symbols

This is a bit situational, but every byte counts!

puts"thing" # before
puts:thing  # after
          ^

Answer 25

Avoid Array#repeated_permutation and Array#repeated_combination

Credit to @AsoneTuhid who golfed the code for repeated permutations of length \$\ge5\$.

Some of Ruby's built-in methods have unfortunately long names. Never use Array#repeated_permutation or Array#repeated_combination; save bytes as follows.

Repeated permutations

Assume a is an array. To get repeated permutations of length \$L = n + 1\$ of the elements of a:

a.product(a,a,a)  # L <= 4; number of arguments = n
a.product(*[a]*n) # L >= 5

Depending on context, the parentheses may not be required. Both of the above yield an array. To iterate over the permutations, simply call with a block.

Repeated combinations

For repeated combinations of length \$L\$, use one of

a.send(a.methods[42],L)    # enumerator
[*a.send(a.methods[42],L)] # array

The index that yields :repeated_combination depends on the Ruby version and possibly the OS (42 is correct for Ruby 2.5.5 on Linux, which is the version on TIO at the time of writing). The default indexing may also be disrupted if any libraries are loaded. The correct index in any case can always be found using [].methods.index(:repeated_combination).

---

In general, calling a method by index using Object#send and Object#methods, as demonstrated above for repeated combinations, is shorter than a direct method call when the number of bytes in the method name and the index of the method in the methods array satisfy:

+-------+-------+
| Bytes | Index |
+-------+-------+
|  19+  |  0-9  |
|  20+  | 10-99 |
|  21+  | 100+  |
+-------+-------+
Subtract 1 from byte count if parentheses not needed for send.

Answer 26

Use Goruby instead of Ruby, which is something like an abbreviated version of Ruby. You can install it with rvm via

rvm install goruby

Goruby allows you to write most of your code as you would be writing Ruby, but has additional abbreviations built in. To find out the shortest available abbreviation for something, you can use the helper method shortest_abbreviation, for example:

shortest_abbreviation :puts
#=> "pts"

Array.new.shortest_abbreviation :map
#=> "m"

String.new.shortest_abbreviation :capitalize
#=> "cp"

Array.new.shortest_abbreviation :join
#=> "j"

Also very handy is the alias say for puts which itself can be abbreviated with s. So instead of

puts [*?a..?z].map(&:capitalize).join

you can now write

s [*?a..?z].m(&:cp).j

to print the alphabet in capitals (which is not avery good example). This blog post explains more stuff and some of the inner workings if you are interested in further reading.

PS: don't miss out on the h method ;-)

Answer 27

I just attempted a TDD code-golf challenge i.e. Write shortest code to make specs pass. The specs were something like

describe PigLatin do
  describe '.translate' do
    it 'translates "cat" to "atcay"' do
      expect(PigLatin.translate('cat')).to eq('atcay')
    end
    # And similar examples for .translate
  end
end

For the sake of code-golf, one need not create a module or class.

Instead of

module PigLatin def self.translate s;'some code'end;end

one can do

def(PigLatin=p).translate s;'some code'end

Saves 13 characters!

Answer 28

When a challenge requires that you output multiple lines, you don't have to loop through your results in order to print each line of e.g. an array. The puts method will flatten an array and print each element on a separate line.

> a = %w(testing one two three)
> puts a
testing
one
two
three

Combining the splat operator with #p you can make it even shorter:

p *a

The splat operator (technically the *@ method, I think) also casts your non-array enumerables to arrays:

> p a.lazy.map{|x|x*2}
#<Enumerator::Lazy: #<Enumerator::Lazy: [1, 2, 3]>:map>

vs

> p *a.lazy.map{|x|x*2}
2
4
6

Answer 29

Subscripting Numbers!

I just discovered this yesterday. n[i] returns n's bit at the i-th position. Example:

irb(main):001:0> n = 0b11010010
=> 210
irb(main):002:0> n[0]
=> 0
irb(main):003:0> n[1]
=> 1
irb(main):004:0> n[2]
=> 0
irb(main):005:0> n[3]
=> 0
irb(main):006:0> n[4]
=> 1
irb(main):007:0> n[5]
=> 0

Answer 30

On looping

while...end

If you need to break out of the loop, while condition;code;end will probably be shorter than loop{code;condition||break}.

The ; before end is not always required, eg. while condition;p("text")end

until c;...;end is equivalent to while !c;...;end and 1 byte shorter.

Note that in most cases code while condition and code until condition are significantly shorter as they don't require the end keyword and can often drop semicolons. Also, i+=1 while true is equivalent to i+=1while true and 1 byte shorter.

redo

When run, the redo command jumps back to the beginning of the block it's in.

When using redo in a lambda, you will have to move any setup variables to the arguments to avoid them being reset at every iteration (see examples).

recursion

Recursion can be shorter is some cases. For instance, if you're working on an array element by element, something like f=->s,*t{p s;t[0]&&f[*t]} can be shorter than the alternatives depending on the stuff.

Note that per the current consensus, if you're calling your function by name, you need to include the assignment (f=) in the byte count making all recursive lambdas 2 bytes longer by default.

eval

If you need to run some code n times, you can use eval"code;"*n.

This will concatenate code; n times and run the whole thing.

Note that in most cases you need to include a ; after your code.

Examples

A lambda to print all numbers from 1 to a inclusive:

->n{i=0;loop{p i+=1;i<n||break}} # 32 bytes
f=->n,i=1{i>n||p(i)&&f[n,i+1]}   # 30 bytes
->n,i=0{p(i+=1)<n&&redo}         # 24 bytes
->n{i=0;p i+=1while i<n}         # 24 bytes
->n{i=0;eval"p i+=1;"*n}         # 24 bytes
->n{n.times{|i|p i+1}}           # 22 bytes # thanks to @benj2240
->n{n.times{p _1+1}}             # 20 bytes # thanks to @AgentIvan

In this case, since the end-point is defined (n), the n.times loop is the shortest.

The redo loop works because i+=1 modifies i and returns its new value and p(x) returns x (this is not true of print and puts).

Given a function g and a number n, find the first number strictly larger than n for which g[n] is truthy

->g,n{loop{g[n+=1]&&break};n}     # 29 bytes
f=->g,n{g[n+=1]?n:f[g,n]}         # 25 bytes
->g,n{1until g[n+=1];n}           # 23 bytes
->g,n{(n+1..).find &g}            # 22 bytes
->g,n{g[n+=1]?n:redo}             # 21 bytes

In this case, with an unknown end-point, redo is the best option.

The (n+1..Inf) loop is equivalent to simply looping indefinitely but more verbose.

A 1 (or anything else) is required before the until keyword to complete the syntax, using a number allows you to drop a space.

The eval method is not viable in this case because there is neither a defined end-point nor an upper bound.

Update: with the new open ranges (n+1..Inf) can be written simply as (n+1..), also .find{|x|g[x]} is equivalent to .find &g where g is converted to a block.

TL;DR check out redo, it can very often shave off a couple of bytes