Should I use `.blank?` in a view of performance?

Question

I understand what .blank? , .nil? and .empty? does now.

I'm thinking why should I not replace all the .nil? and .empty? to .blank? for less risks of making mistakes. For example, if current_user.blank?

Is there a performance issue for .blank? method? Is it slower or consuming more memory?

If it is, how bad is it? As bad as string vs symbol?

Thanks in advance.

Its worrying you are that inconfident in your coding skills. I dont see any reason to always use .blank? — apneadiving, Dec 28 '13 at 11:52
`.nil?` returns `true` is instance is `nil`, `.empty?` returns `true` when instance, like array or string is empty. and `blank?` should return `true` if nil or empty. — Малъ Скрылевъ, Dec 28 '13 at 11:53
`current_user.blank?` is a tad overkill, it doesn't make sense for a user to be empty. — Alexei Averchenko, Dec 28 '13 at 12:03
Sorry for misleading title. It must be a very common fact that everyone answers from title in SO. — Ivan Wang, Dec 28 '13 at 18:03
I mean not only title but and the context of the message. What is the performance in your meaning? — Малъ Скрылевъ, Dec 28 '13 at 19:51
@IvanWang Sorry, for the harsh word... I've changed my post. — Малъ Скрылевъ, Dec 28 '13 at 20:15
It's okay. I have got the answer any way. I was over-thinking how .blank? is processed in rails. — Ivan Wang, Dec 29 '13 at 08:36

score 6 · Accepted Answer · edited May 23 '17 at 12:08

Theory

All three methods have various realizations for various classes:

:nil?:

# NilClass
def nil?
  true
end

# Object
def nil?
  false
end

:empty?:

# String, Array
def empty?
  if self.size == 0
    true
  else
    false
  end
end

:blank?:

# FalseClass, NilClass
def blank?
  true
end

# Object
def blank?
  respond_to?(:empty?) ? empty? : !self
end

# TrueClass
def blank?
  false
end

# String
def blank?
  self !~ /[^[:space:]]/
end

As you may see the various classes implement various methods style. In case of String class it takes time of a single Regexp, in case of Object, including Hash, and Array it takes time of call to :respond and return a value nil or not Object. The seconds are just operations that takes time similar to :nil?. :respond? method checks the presense of the :empty? method that takes theoretically slight more times than two times to :empty?.

Investigations

I wrote simple script that simulates the behaviour of those methods, and calculates execution time of them:

#! /usr/bin/env ruby

require 'benchmark'

obj = Object.new
array = []
empty_string = ''
non_empty_string = '   '

funcs = 
[ proc { empty_string.empty? }, 
  proc { non_empty_string.empty? },
  proc { obj.nil? },
  proc { nil.nil? },
  proc { true },
  proc { respond_to?(:empty?) ? empty? : !self },
  proc { array.respond_to?(:empty?) ? array.empty? : !array },
  proc { non_empty_string !~ /[^[:space:]]/ } ]

def ctime func
   time = 0
   1000.times { time += Benchmark.measure { 1000.times { func.call } }.to_a[5].to_f }
   rtime = time /= 1000000
end

funcs.each {| func | p ctime( func ) }

And results:

# empty String :empty?
# 4.604020118713379e-07
# non_empty String :empty?
# 4.5903921127319333e-07
# Object :nil?
# 5.041143894195557e-07
# NilClass :nil?
# 4.7951340675354e-07
# FalseClass, NilClass, TrueClass :blank?
# 4.09862756729126e-07
# main :blank? ( respond_to returns false )
# 6.444177627563477e-07
# Array :blank? ( respond_to returns true )
# 6.491720676422119e-07
# String :blank?
# 1.4315705299377442e-06

As you may see, obvious champion from the end of table in case of speed is method :blank? of String class. It has execution time descreased in 3-4 times against an simple empty? method of a class. Object's :blank? method has only 1.5 times execution time degradation. Note, that :respond_to method in it has just a few time to execute, becase as far as I see the ruby interpreter caches the result of its execution. So, conclusion. Try to avoid using String's .blank? method.

Additional, if you need to know how to use the methods see here.

This is a great full detail report of everything! Thank you so much. Learnt much more than I have expected. Right, `.blank?` is doing too perfect on checking white space string. But not so perfect on checking array. Really interesting. — Ivan Wang, Dec 29 '13 at 18:08
@IvanWang Vise versa, `:blank?` on `String` takes more time then the same on `Array` or `Hash`. so on `String` it becomes worse more than twice. — Малъ Скрылевъ, Dec 29 '13 at 18:13
I have confused you again. I meant `.blank?` is doing literally perfect on string, because it checks every white space in string. You are right, doing too perfect surely will make it slow. However, `.blank?` decided to not do white space checking on array. That makes it seems "half-way-done". (I think that is for a good reason. If it does white space check on every instant inside an array, it will take ages if the array is full of white space instances.) — Ivan Wang, Dec 29 '13 at 18:28
Yes, functionally `:blank?` for `String` is better, then `Array`'s one. May be it because of that `Array` `[nil]`, or `Hash` `{ nil => nil }` aren't frequently occurred in Rails =) — Малъ Скрылевъ, Dec 29 '13 at 18:32

Siva · Answer 2 · 2013-12-28T12:11:25.183

3

check the source here blank?

def blank?
  respond_to?(:empty?) ? empty? : !self
end

Technically its a 2x of empty?

is there a performance issue for .blank? ? Is it slower or consuming more memory?

Well, that depends on lot of factors

amount of data your app handles
server hardware
no of incoming requests

But frankly this is not gonna bother you unless you run twitter

edited Dec 28 '13 at 12:11

answered Dec 28 '13 at 12:05

Siva

7,780
6
47
54

Nice. The source codes help me understand much better. 2X .empty? sounds reasonable. Now I can just feel free to use more .blank on small scale projects. – Ivan Wang Dec 28 '13 at 12:31

Should I use `.blank?` in a view of performance?

2 Answers2

Theory

Investigations