I have a situation where merge! seems to modify the value of a CONST. Can this occur? How?
I'm doing some API ingestion and mapping, like you do...
module Placement
FEATURE_DEFAULTS = {
"thingone" => "false",
"thingtwo" => "false"
}
def extract_features!(feat)
feat['norm_features'] ||= FEATURE_DEFAULTS
feat['norm_features'].merge!(
Array(feat['attributes']['feature']).reduce({}) do |h,f|
h[f] = "true"
h
end
)
end
def get_placement(_opts)
data_source["things"]["thing"].map do |thing|
product = {}
thing.each do |key, value|
new_key = RENAME_FIELDS[key] || key
new_value = REPLACE_FIELDS[key] || value
product[new_key] = new_value
end
binding.pry # 1
extract_features!(product)
binding.pry # 2
product
end
end
end
Later, I include this in a class for the API client, and then I call the get_placement method.
Dilemma
step 1
For the first run, in pry binding 1 & 2, the value of FEATURE_DEFAULTS is as seen above. For 2 the value of FEATURE_DEFAULTS is the same, and the value of product['norm_features'] is the same (plus the results of the meshing operation of extract_features!
step 2
The output (to the caller of get_placement), for every thing/product, is
FEATURE_DEFAULTS = {
"thingone" => true,
"thingtwo" => true
}
step 3
When I run this the second time (after starting up the service / app), the value of FEATURE_DEFAULTS, and pry binding 1 & 2, is
FEATURE_DEFAULTS = {
"thingone" => true,
"thingtwo" => true
}
What is happening here?
This seems to confirm that, after running the extract_features! method, the FEATURE_DEFAULTS CONST is changed. If I do not use merge! in extract_features!, and use merge instead, then the CONST value does not change.
I can post more code, if needed, or most to a Gist.
Ruby MRI 2.2.2
I am doing this inside a rails app, but I don't see that it matters.
I have a situation where merge! seems to modify the value of a CONST. Can this occur? How?
No, in general, methods cannot change variable bindings, regardless of whether those variables are local variables, instance variables, class variables, global variables, or constants. Variables aren't objects in Ruby, you can't call methods on them, you can't pass them as arguments to methods, ergo, you can't tell them to change themselves.
The exception to this are meta-programming methods like Binding#local_variable_set, Object#instance_variable_set, Module#class_variable_set, or Module#const_set.
What you can do, however, and what you are doing in this case, is tell the object the variable points to to change itself. The documentation for Hash#merge! is unfortunately a bit unclear in that it does not explicitly mention that Hash#merge! mutates its receiver.
This seems to confirm that, after running the extract_features! method, the FEATURE_DEFAULTS CONST is changed. If I do not use merge! in extract_features!, and use merge instead, then the CONST value does not change.
No, it only confirms that the state of the object FEATURE_DEFAULTS points to changed, it doesn't say anything about whether the binding of FEATURE_DEFAULTS, i.e. which object it points to, changed. You can confirm that the constant still points to the same object by looking at its object_id.
Of course, constants can be re-assigned anyway, albeit triggering a warning.
I want to change the default arguments passed to a Ruby function. For example, instead of each time writing
[1,2,3].do_stuff(:option => ' my option ')
I want to modify the defaults so that I can write
[1,2,3].do_stuff
What is the simplest, cleanest, most Ruby-like way of changing default parameters?
>> [1, 2, 3].do_stuff
=> Result I get
>> [1, 2, 3].do_stuff :an_option => a_value
=> Result I really want, but don't want to specify the argument
I like to use super for this. It allows us to add some functionality to the method apart from just changing default arguments:
class Array
def do_stuff(options = {})
# Verify if caller has not passed the option
options[:argument_i_want_to_change] = default_value_i_want unless options.has_key? :argument_i_want_to_change
# call super
super
end
end
Result:
>> [1, 2, 3].do_stuff
=> Result that I really want
UPDATE: Removed reverse_merge! dependency. (Now looking for a better alternatives to using []= method)
(moved from your original question)
I assume you are talking about a method Array#do_stuff that already exists, but you want to modify it slightly (in your case by changing a default parameter).
A post here gives a nice way of doing it. It doesn't suffer from the same problems as the alias technique, as there isn't a leftover "old" method.
Here how you could use that technique with your example problem (tested with ruby 1.9)
class Array
old_do_stuff = instance_method(:do_stuff)
define_method(:do_stuff) { |options = {}|
options[:option] ||= " option "
old_do_stuff.bind(self).call(options)
}
end
You might also want read up on UnboundMethod if the above code is confusing. Note that old_do_stuff goes out of scope after the end statement, so it isn't a problem for future uses of Array.
Are you wanting a solution for code you didn't write yourself? There are two options I'm aware of.
Code you wrote yourself:
def some_method_you_wrote(options)
becomes:
def some_method_you_wrote(options = { :option1 => 'value' })
(Swanand's answer is nice too)
For code you didn't write, look into aliasing methods. (Rails provides something called alias_method_chain for this purpose, IIRC.)
I have a a hash
foo = {'bar'=>'baz'}
I would like to call foo.bar #=> 'baz'
My motivation is rewriting an activerecord query into a raw sql query (using Model#find_by_sql). This returns a hash with the SELECT clause values as keys. However, my existing code relies on object.method dot notation. I'd like to do minimal code rewrite. Thanks.
Edit: it appears Lua has this feature:
point = { x = 10, y = 20 } -- Create new table
print(point["x"]) -- Prints 10
print(point.x) -- Has exactly the same meaning as line above
>> require 'ostruct'
=> []
>> foo = {'bar'=>'baz'}
=> {"bar"=>"baz"}
>> foo_obj = OpenStruct.new foo
=> #<OpenStruct bar="baz">
>> foo_obj.bar
=> "baz"
>>
What you're looking for is called OpenStruct. It's part of the standard library.
A good solution:
class Hash
def method_missing(method, *opts)
m = method.to_s
if self.has_key?(m)
return self[m]
elsif self.has_key?(m.to_sym)
return self[m.to_sym]
end
super
end
end
Note: this implementation has only one known bug:
x = { 'test' => 'aValue', :test => 'bar'}
x.test # => 'aValue'
If you prefer symbol lookups rather than string lookups, then swap the two 'if' condition
Rather than copy all the stuff out of the hash, you can just add some behaviour to Hash to do lookups.
If you add this defintion, you extend Hash to handle all unknown methods as hash lookups:
class Hash
def method_missing(n)
self[n.to_s]
end
end
Bear in mind that this means that you won't ever see errors if you call the wrong method on hash - you'll just get whatever the corresponding hash lookup would return.
You can vastly reduce the debugging problems this can cause by only putting the method onto a specific hash - or as many hashes as you need:
a={'foo'=>5, 'goo'=>6}
def a.method_missing(n)
self[n.to_s]
end
The other observation is that when method_missing gets called by the system, it gives you a Symbol argument. My code converted it into a String. If your hash keys aren't strings this code will never return those values - if you key by symbols instead of strings, simply substitute n for n.to_s above.
There are a few gems for this. There's my recent gem, hash_dot, and a few other gems with similar names I discovered as I released mine on RubyGems, including dot_hash.
HashDot allows dot notation syntax, while still addressing concerns about NoMethodErrors addressed by #avdi. It is faster, and more traversable than an object created with OpenStruct.
require 'hash_dot'
a = {b: {c: {d: 1}}}.to_dot
a.b.c.d => 1
require 'open_struct'
os = OpenStruct.new(a)
os.b => {c: {d: 1}}
os.b.c.d => NoMethodError
It also maintains expected behavior when non-methods are called.
a.non_method => NoMethodError
Please feel free to submit improvements or bugs to HashDot.
What's an efficient, rubyesque way of testing if a collection supports string indexing?
Long version:
I'd like for a class of mine to normalize certain values. To achieve this, an instance takes a collection of values as its 'values' attribute. I'd like values= to accept both lists (integer indexed collections, including the built-in Array) and associative arrays (object indexed collections, such as Hash). If passed a list, it turns the list into an associative array by inverting keys & values. This means the method needs to distinguish lists from associative arrays. It should also work on any indexed collection, not just descendants of Array and Hash, so any sort of type sniffing the collection type is considered ugly and wrong. Type sniffing the index type, however...
Currently, I'm using exceptions to tell the difference, but I prefer to use exceptions for, well, exceptional circumstances rather than a general control structure. It's just a personal preference, one I'm not too attached to. If exceptions are the ruby way to solve this problem, please let me know.
def values=(values)
begin
values['']
#values = values.dup
rescue TypeError
#values = Hash[ values.zip((0..values.length-1).to_a) ]
end
#values.each_value { |v| #values[v] = v}
end
Note: a complete solution would take the transitive closure of values, but for now I can assume the keys & values of values are from different domains.
The point of all this is to enable code like:
toggle.values = [:off, :on]
toggle.normalise(:off) == 0
toggle.normalise(1) == 1
bool.values = {:off => 0, :false => 0, :no => 0,
:on => 1, :true => 1, :yes => 1}
bool.normalise(:yes) == 1
bool.normalise(0) == 0
PS. This is for a personal project, so elegance and the Ruby way are paramount. I'm looking for interesting approaches, especially if they illustrate an interesting concept (such as "exceptions can be used as behavioral tests").
Duck typing to the rescue!
hash = if collection.respond_to? :to_hash
collection.to_hash
elsif collection.respond_to? :to_ary
collection.to_ary.inject({}) { |_hash,(key,value)| _hash.merge!(key => value) }
else if collection.respond_to? :inject
collection.inject({}) { |_hash,(key,value)| _hash.merge!(key => value) }
else
raise ArgumentError, "not a collection type I understand"
end
if want_dupe and collection.object_id == hash.object_id
hash = hash.dup
end
Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
Continuing the "Hidden features of ..." meme, let's share the lesser-known but useful features of Ruby programming language.
Try to limit this discussion with core Ruby, without any Ruby on Rails stuff.
See also:
Hidden features of C#
Hidden features of Java
Hidden features of JavaScript
Hidden features of Ruby on Rails
Hidden features of Python
(Please, just one hidden feature per answer.)
Thank you
From Ruby 1.9 Proc#=== is an alias to Proc#call, which means Proc objects can be used in case statements like so:
def multiple_of(factor)
Proc.new{|product| product.modulo(factor).zero?}
end
case number
when multiple_of(3)
puts "Multiple of 3"
when multiple_of(7)
puts "Multiple of 7"
end
Peter Cooper has a good list of Ruby tricks. Perhaps my favorite of his is allowing both single items and collections to be enumerated. (That is, treat a non-collection object as a collection containing just that object.) It looks like this:
[*items].each do |item|
# ...
end
Don't know how hidden this is, but I've found it useful when needing to make a Hash out of a one-dimensional array:
fruit = ["apple","red","banana","yellow"]
=> ["apple", "red", "banana", "yellow"]
Hash[*fruit]
=> {"apple"=>"red", "banana"=>"yellow"}
One trick I like is to use the splat (*) expander on objects other than Arrays. Here's an example on a regular expression match:
match, text, number = *"Something 981".match(/([A-z]*) ([0-9]*)/)
Other examples include:
a, b, c = *('A'..'Z')
Job = Struct.new(:name, :occupation)
tom = Job.new("Tom", "Developer")
name, occupation = *tom
Wow, no one mentioned the flip flop operator:
1.upto(100) do |i|
puts i if (i == 3)..(i == 15)
end
One of the cool things about ruby is that you can call methods and run code in places other languages would frown upon, such as in method or class definitions.
For instance, to create a class that has an unknown superclass until run time, i.e. is random, you could do the following:
class RandomSubclass < [Array, Hash, String, Fixnum, Float, TrueClass].sample
end
RandomSubclass.superclass # could output one of 6 different classes.
This uses the 1.9 Array#sample method (in 1.8.7-only, see Array#choice), and the example is pretty contrived but you can see the power here.
Another cool example is the ability to put default parameter values that are non fixed (like other languages often demand):
def do_something_at(something, at = Time.now)
# ...
end
Of course the problem with the first example is that it is evaluated at definition time, not call time. So, once a superclass has been chosen, it stays that superclass for the remainder of the program.
However, in the second example, each time you call do_something_at, the at variable will be the time that the method was called (well, very very close to it)
Another tiny feature - convert a Fixnum into any base up to 36:
>> 1234567890.to_s(2)
=> "1001001100101100000001011010010"
>> 1234567890.to_s(8)
=> "11145401322"
>> 1234567890.to_s(16)
=> "499602d2"
>> 1234567890.to_s(24)
=> "6b1230i"
>> 1234567890.to_s(36)
=> "kf12oi"
And as Huw Walters has commented, converting the other way is just as simple:
>> "kf12oi".to_i(36)
=> 1234567890
Hashes with default values! An array in this case.
parties = Hash.new {|hash, key| hash[key] = [] }
parties["Summer party"]
# => []
parties["Summer party"] << "Joe"
parties["Other party"] << "Jane"
Very useful in metaprogramming.
Another fun addition in 1.9 Proc functionality is Proc#curry which allows you to turn a Proc accepting n arguments into one accepting n-1. Here it is combined with the Proc#=== tip I mentioned above:
it_is_day_of_week = lambda{ |day_of_week, date| date.wday == day_of_week }
it_is_saturday = it_is_day_of_week.curry[6]
it_is_sunday = it_is_day_of_week.curry[0]
case Time.now
when it_is_saturday
puts "Saturday!"
when it_is_sunday
puts "Sunday!"
else
puts "Not the weekend"
end
Download Ruby 1.9 source, and issue make golf, then you can do things like this:
make golf
./goruby -e 'h'
# => Hello, world!
./goruby -e 'p St'
# => StandardError
./goruby -e 'p 1.tf'
# => 1.0
./goruby19 -e 'p Fil.exp(".")'
"/home/manveru/pkgbuilds/ruby-svn/src/trunk"
Read the golf_prelude.c for more neat things hiding away.
Boolean operators on non boolean values.
&& and ||
Both return the value of the last expression evaluated.
Which is why the ||= will update the variable with the value returned expression on the right side if the variable is undefined. This is not explicitly documented, but common knowledge.
However the &&= isn't quite so widely known about.
string &&= string + "suffix"
is equivalent to
if string
string = string + "suffix"
end
It's very handy for destructive operations that should not proceed if the variable is undefined.
The Symbol#to_proc function that Rails provides is really cool.
Instead of
Employee.collect { |emp| emp.name }
You can write:
Employee.collect(&:name)
One final one - in ruby you can use any character you want to delimit strings. Take the following code:
message = "My message"
contrived_example = "<div id=\"contrived\">#{message}</div>"
If you don't want to escape the double-quotes within the string, you can simply use a different delimiter:
contrived_example = %{<div id="contrived-example">#{message}</div>}
contrived_example = %[<div id="contrived-example">#{message}</div>]
As well as avoiding having to escape delimiters, you can use these delimiters for nicer multiline strings:
sql = %{
SELECT strings
FROM complicated_table
WHERE complicated_condition = '1'
}
Use a Range object as an infinite lazy list:
Inf = 1.0 / 0
(1..Inf).take(5) #=> [1, 2, 3, 4, 5]
More info here: http://banisterfiend.wordpress.com/2009/10/02/wtf-infinite-ranges-in-ruby/
I find using the define_method command to dynamically generate methods to be quite interesting and not as well known. For example:
((0..9).each do |n|
define_method "press_#{n}" do
#number = #number.to_i * 10 + n
end
end
The above code uses the 'define_method' command to dynamically create the methods "press1" through "press9." Rather then typing all 10 methods which essentailly contain the same code, the define method command is used to generate these methods on the fly as needed.
module_function
Module methods that are declared as module_function will create copies of themselves as private instance methods in the class that includes the Module:
module M
def not!
'not!'
end
module_function :not!
end
class C
include M
def fun
not!
end
end
M.not! # => 'not!
C.new.fun # => 'not!'
C.new.not! # => NoMethodError: private method `not!' called for #<C:0x1261a00>
If you use module_function without any arguments, then any module methods that comes after the module_function statement will automatically become module_functions themselves.
module M
module_function
def not!
'not!'
end
def yea!
'yea!'
end
end
class C
include M
def fun
not! + ' ' + yea!
end
end
M.not! # => 'not!'
M.yea! # => 'yea!'
C.new.fun # => 'not! yea!'
Short inject, like such:
Sum of range:
(1..10).inject(:+)
=> 55
Warning: this item was voted #1 Most Horrendous Hack of 2008, so use with care. Actually, avoid it like the plague, but it is most certainly Hidden Ruby.
Superators Add New Operators to Ruby
Ever want a super-secret handshake operator for some unique operation in your code? Like playing code golf? Try operators like
-~+~-
or
<---
That last one is used in the examples for reversing the order of an item.
I have nothing to do with the Superators Project beyond admiring it.
I'm late to the party, but:
You can easily take two equal-length arrays and turn them into a hash with one array supplying the keys and the other the values:
a = [:x, :y, :z]
b = [123, 456, 789]
Hash[a.zip(b)]
# => { :x => 123, :y => 456, :z => 789 }
(This works because Array#zip "zips" up the values from the two arrays:
a.zip(b) # => [[:x, 123], [:y, 456], [:z, 789]]
And Hash[] can take just such an array. I've seen people do this as well:
Hash[*a.zip(b).flatten] # unnecessary!
Which yields the same result, but the splat and flatten are wholly unnecessary--perhaps they weren't in the past?)
Auto-vivifying hashes in Ruby
def cnh # silly name "create nested hash"
Hash.new {|h,k| h[k] = Hash.new(&h.default_proc)}
end
my_hash = cnh
my_hash[1][2][3] = 4
my_hash # => { 1 => { 2 => { 3 =>4 } } }
This can just be damn handy.
Destructuring an Array
(a, b), c, d = [ [:a, :b ], :c, [:d1, :d2] ]
Where:
a #=> :a
b #=> :b
c #=> :c
d #=> [:d1, :d2]
Using this technique we can use simple assignment to get the exact values we want out of nested array of any depth.
Class.new()
Create a new class at run time. The argument can be a class to derive from, and the block is the class body. You might also want to look at const_set/const_get/const_defined? to get your new class properly registered, so that inspect prints out a name instead of a number.
Not something you need every day, but quite handy when you do.
create an array of consecutive numbers:
x = [*0..5]
sets x to [0, 1, 2, 3, 4, 5]
A lot of the magic you see in Rubyland has to do with metaprogramming, which is simply writing code that writes code for you. Ruby's attr_accessor, attr_reader, and attr_writer are all simple metaprogramming, in that they create two methods in one line, following a standard pattern. Rails does a whole lot of metaprogramming with their relationship-management methods like has_one and belongs_to.
But it's pretty simple to create your own metaprogramming tricks using class_eval to execute dynamically-written code.
The following example allows a wrapper object to forwards certain methods along to an internal object:
class Wrapper
attr_accessor :internal
def self.forwards(*methods)
methods.each do |method|
define_method method do |*arguments, &block|
internal.send method, *arguments, &block
end
end
end
forwards :to_i, :length, :split
end
w = Wrapper.new
w.internal = "12 13 14"
w.to_i # => 12
w.length # => 8
w.split('1') # => ["", "2 ", "3 ", "4"]
The method Wrapper.forwards takes symbols for the names of methods and stores them in the methods array. Then, for each of those given, we use define_method to create a new method whose job it is to send the message along, including all arguments and blocks.
A great resource for metaprogramming issues is Why the Lucky Stiff's "Seeing Metaprogramming Clearly".
use anything that responds to ===(obj) for case comparisons:
case foo
when /baz/
do_something_with_the_string_matching_baz
when 12..15
do_something_with_the_integer_between_12_and_15
when lambda { |x| x % 5 == 0 }
# only works in Ruby 1.9 or if you alias Proc#call as Proc#===
do_something_with_the_integer_that_is_a_multiple_of_5
when Bar
do_something_with_the_instance_of_Bar
when some_object
do_something_with_the_thing_that_matches_some_object
end
Module (and thus Class), Regexp, Date, and many other classes define an instance method :===(other), and can all be used.
Thanks to Farrel for the reminder of Proc#call being aliased as Proc#=== in Ruby 1.9.
The "ruby" binary (at least MRI's) supports a lot of the switches that made perl one-liners quite popular.
Significant ones:
-n Sets up an outer loop with just "gets" - which magically works with given filename or STDIN, setting each read line in $_
-p Similar to -n but with an automatic puts at the end of each loop iteration
-a Automatic call to .split on each input line, stored in $F
-i In-place edit input files
-l Automatic call to .chomp on input
-e Execute a piece of code
-c Check source code
-w With warnings
Some examples:
# Print each line with its number:
ruby -ne 'print($., ": ", $_)' < /etc/irbrc
# Print each line reversed:
ruby -lne 'puts $_.reverse' < /etc/irbrc
# Print the second column from an input CSV (dumb - no balanced quote support etc):
ruby -F, -ane 'puts $F[1]' < /etc/irbrc
# Print lines that contain "eat"
ruby -ne 'puts $_ if /eat/i' < /etc/irbrc
# Same as above:
ruby -pe 'next unless /eat/i' < /etc/irbrc
# Pass-through (like cat, but with possible line-end munging):
ruby -p -e '' < /etc/irbrc
# Uppercase all input:
ruby -p -e '$_.upcase!' < /etc/irbrc
# Same as above, but actually write to the input file, and make a backup first with extension .bak - Notice that inplace edit REQUIRES input files, not an input STDIN:
ruby -i.bak -p -e '$_.upcase!' /etc/irbrc
Feel free to google "ruby one-liners" and "perl one-liners" for tons more usable and practical examples. It essentially allows you to use ruby as a fairly powerful replacement to awk and sed.
The send() method is a general-purpose method that can be used on any Class or Object in Ruby. If not overridden, send() accepts a string and calls the name of the method whose string it is passed. For example, if the user clicks the “Clr” button, the ‘press_clear’ string will be sent to the send() method and the ‘press_clear’ method will be called. The send() method allows for a fun and dynamic way to call functions in Ruby.
%w(7 8 9 / 4 5 6 * 1 2 3 - 0 Clr = +).each do |btn|
button btn, :width => 46, :height => 46 do
method = case btn
when /[0-9]/: 'press_'+btn
when 'Clr': 'press_clear'
when '=': 'press_equals'
when '+': 'press_add'
when '-': 'press_sub'
when '*': 'press_times'
when '/': 'press_div'
end
number.send(method)
number_field.replace strong(number)
end
end
I talk more about this feature in Blogging Shoes: The Simple-Calc Application
Fool some class or module telling it has required something that it really hasn't required:
$" << "something"
This is useful for example when requiring A that in turns requires B but we don't need B in our code (and A won't use it either through our code):
For example, Backgroundrb's bdrb_test_helper requires 'test/spec', but you don't use it at all, so in your code:
$" << "test/spec"
require File.join(File.dirname(__FILE__) + "/../bdrb_test_helper")
Defining a method that accepts any number of parameters and just discards them all
def hello(*)
super
puts "hello!"
end
The above hello method only needs to puts "hello" on the screen and call super - but since the superclass hello defines parameters it has to as well - however since it doesn't actually need to use the parameters itself - it doesn't have to give them a name.
private unless Rails.env == 'test'
# e.g. a bundle of methods you want to test directly
Looks like a cool and (in some cases) nice/useful hack/feature of Ruby.