Arrays in Ruby: Benefits and Costs to Use Them

Data structures are eagerly introduced to new programmers, since they will be using them pretty much through their whole career. Knowing them, how they work, what are they good for, which one to use, are things that can change the performance of your solutions quite vastly. On this article, as the title suggest, I will be talking arrays with ruby, and going into a bit of detail here and there about how this language manages them. Article is intended to beginners with a general understanding of basic concepts . Some ruby array knowledge is necessary to follow some examples, such as adding elements, concatenation, and looping. Arrays. What, how and where. Arrays are, , a collection of objects. This definition is quite literal in most languages, meaning that objects stored in said structure would be placed sequentially in memory as well as in the abstraction that the developer will be using at a high level. as the word semantic suggest consecutive Usually, we would find two different 'types' of Arrays. Static, and Dynamic. Static arrays are those which we define the size, and that is set in stone. Dynamic arrays, in the other hand, grow or shrink, typically used when we do not know the final size of this structure. Check this next example of an implementation (using C not ruby, but bear with me) { intArray [ ]; ( , ( )); ( i= ; i int main () int 10 printf "Size of int: %ld bytes\n" sizeof int for int 0 10 printf "\nElement\t%p" return 0 Now, this trivial piece of code, when executed will output a very literal representation of what an array in a programming a computer is. usually Size of : bytes Element Element Element Element Element Element Element Element Element Element int 4 0x7ffe9a634b50 0x7ffe9a634b54 0x7ffe9a634b58 0x7ffe9a634b5c 0x7ffe9a634b60 0x7ffe9a634b64 0x7ffe9a634b68 0x7ffe9a634b6c 0x7ffe9a634b70 0x7ffe9a634b74 More specifically, a static array that We can see that every integer, will have up to 4 bytes of data and even thought our array is still empty, memory is already reserved an ready to use. We defined a static array of integers. will hold up to 10 integers. Not only that, but their address are separated by these 4 bytes in particular, which gives you an idea of how it could work. You have the first memory cell address, if you go 4 bytes from there, you will find the second element. So, we can assume that something like this : [ ] array 4 could work by starting on the first direction and adding 4*[size of object] to that address, to get to that direction and get the value in it. Why is it important Basically, because the first concept of " ". Therefore, you first need to find a place in memory with [size] consecutive free spaces before start placing them. that memory is already reserved ? everything is next to each other in memory " ". That is correct! And that is because you were using a But wait, I always use .push(element) in Ruby and never had an issue with reserving memory beforehand dynamic array. Dynamic Arrays A dynamic array is a particular structure that, although still being an array, is a bit more complex and has some more functionality added to make your life easier by making the everyday array feel extremely flexible. For now on I can use Ruby. And don't worry, no more C. Let's start by writing a few plain examples: array = [] .times array.push(x) p array p +array.length().to_s prefixSumArray = [] array.each prefixSumArray.push(x+ ) arrayCopy = array array.concat(arrayCopy.concat(prefixSumArray)) p array p +array.length().to_s 5 do |x| end "First array length: " do |x| 5 end "First array length after... playing: " All right all right. That's a lot of action. Lets break it down. array = [] Defining a new array with 0 elements, therefore, from i = 0 to... well 0. On the first loop, we are adding 5 elements with .push method. And let's stop there. We know that array.push adds an element, and if this array did not have that original length then, it would make it bigger. Now that you have 1 element, on an originally 0 sized array, your memory reserved from this array is for one element. Right? Well, no. Since we are working with dynamics array, Ruby (and usually every other approach) have some for that you explicitly say that your array will have. extra space reserved every element So right now, your memory block for this array would look something like this: Adding one more, would be Next step, And by this point you might have already noticed, that when we end our new dynamic array after all this push and concat and adding, we will have quite more [0][]. [0][1]. [0][1][2][][][]. memory reserved than we wanted to use, or that we ended up using. On average can waste up to O(n) . But you might use all of it after all ! " ?" Good question, and basically is just a trade. You are trading for You are making your code faster, by using a bit more space to help you. What goes into every push for a new element, outside the original range of the array, it can be quite expensive ; Oh no! But I did not wanted to waste space ! Why would they do that space complexity time complexity. At the very least you will be going through the memory and look for a place where your array will fit. And this involves looking and reserving again, plus some cleaning of memory that you wont be using any more. With all this in mind, every time you are using a built-in Ruby method that involves expanding an array, what is happening under the hood might be actually that. So that one-liner may not be the best for solving problems where time is crucial or you have to manage big chunks of data. Arrays & Hash. I noticed, learning about the subject, that every time you get a new tool you overuse it a bit. Kind of like a new toy. So a lot of people tend to use Hash when there is really no need for it. (I've been there. And I've seen a lot of people there). For instance, if you need to map objects like with 0..25, you can easily solve that in... well, infinite ways. But lets check a few obvious ones just for the sake of benchmarking. { name: string, telephone: integer } ; hash_test = Hash.new() .times { hash_test[x] = +x } hash_test array_test = [] .times { array_test.push(x+ ) } array_test array_test = Array.new( ) .times { array_test[x] = x+ } array_test Benchmark.bm x.report { doByHash } x.report { doByPush } x.report { pseudoAlloc } # Your code here! require 'benchmark' def doByHash 100001 |x| 1000000 end def doByPush 100001 |x| 100000 end def pseudoAlloc 100001 100001 |x| 100000 end do |x| end You can run an implementation here . But I highly recommend to benchmark it on your local computer so server connection does not affect the result. Also run it several times to get an average and b eware of the numbers that you try to test !! The average total time results that I got for n = 100k was : The amazing with : hash-man 0.026240 # hash Your friendly neighbour the dynamic array that will eventually identify himself as a linked list : 0.014025 # array with push The edgy teen who still is a dynamic array: 0.012692 # array with init(size) RAW_SINGLE_OUTPUT user system total real 0.031000 0.000000 0.031000 ( 0.025804) 0.015000 0.000000 0.015000 ( 0.015524) 0.000000 0.000000 0.000000 ( 0.012109) Code is just stating the obvious in terms of speed in are specially different in behaviour when it comes to this, because theoretically each node can be anywhere in memory. And, as we can see, given that Ruby, as far as my lectures goes, do not have static arrays, even if you define size by default the speed difference that you will get is not a lot that if you did not bother. this particular situation. Hash Keep in mind that this will grow over data size. Sometimes, you will be manipulating data in a time-sensitive code. Reducing complexity of your algorithm is out of the table since you are already doing the best it can. What is left? Hardcode optimizing. And this little trick is one of a few that I've been encountering in my journey with Ruby learning. Closing remarks on Arrays. A bit on Statics: To sum up this bit of data. When we are thinking , we have that cute advantage of " ". Only if you know you will be using that. Because once you defined array, . No deleting, no inserting. Will not get bigger, and will not get smaller. Unless you delete it, and do it all over again. static arrays no wasted space And, that is great ! that that's it A byte on dynamics: When we are using dynamic arrays, we will be potentially wasting space and a bit of time here an there. But also you will gain so much flexibility, that I am sure you will be jealous of people who wrote that source code for having so much fun. , are now way easier for you to use. Same index complexity than his static array brother and O(n) of wasted space here and there is your only price for those extra functions. Since you are using Ruby, I am assuming, you will not think about the edgy static array too much and all that nonsense about memory. But now you know that from time to time, that last push(number), might have caused quite a bit of work on that . Inserts & deletes harmless array of yours Conclusion In this article we talked a little about what dynamic & static arrays are. And a quick pick on what are they doing. Now, please keep in mind that I've been bashing on hashes, on particular situation. And if you paid attention, I said that three times. ( ) There is a reason why hashes are EXTREMELY useful. That of course is out of the scope of this article, but the point remain. For more information about Ruby there is no better source than . Don't be afraid to at the source code of every function to know what is going on just a glance thought... people who go in there, usually don't come back. THIS this particular situation. 4 Make sure you always have more than hammers to use, or everything will look like a nail ! themselves glance Thank you for stopping by!