I pretty much exclusively use Kotlin for competitive programming, mostly because it's the language I'm currently most comfortable with. Here are some scattered notes and tidbits about my experience which I think might be useful to others; if you have any tips/suggestions, feel free to let me know.
Primer
- Kotlin has an official primer for competitive programming. However, the IO code suggested there is only so-so; it's definitely better than
Scanner
, but you definitely can save a lot of runtime in heavy input problems by using the classic Java combination ofBufferedReader
+StringTokenizer
Useful features
A lot less boilerplate than Java. Members are
public
by default. Type inference means a lot less "Pokémon speak". Variables and functions can be declared straight in the top-level of the file. (basically the equivalent ofstatic
functions). Fields have implicit getters and setters that can easily be overridden when necessary.PHP-like string templates, e.g.
"$id $cost"
Extension functions – syntactic sugar for static functions; gives more natural "afterthought" syntax, as well as allowing direct access to public members of the receiver
data class
es – basically custom tuples. Allows convenient destructuring declarations too.Has access to the data structures in the Java standard library (
TreeMap
,HashMap
,PriorityQueue
etc.), and also can useBigInteger
andBigDecimal
if neededFunctional idioms for collection manipulation –
map
,fold
,filter
, etc.Sequences – lazy sequence generators, potentially infinite, has standard collection manipulation functions too. Using the
sequence { ... }
block function allows building a sequence using a scopedyield(value)
function, reminiscent of Pythoninline class
es – allows the creation of a new type that wraps over a base type, but that is represented by an underlying type at runtime. Especially useful for "modulo $$$10^9 + 7$$$" problems, as I keep code for aModInt
class that overloads the arithmetic operators appropriately, but is represented as a plainint
in JVM runtime. Keep in mind that they are experimental as of Kotlin 1.3, but that's fine for CP in my opinionunsigned integer types in the standard library that use the
inline class
feature. Not used very often, but handy if neededinline fun
ctions – tells the compiler to inline the function to call sites. Useful for higher-order functions (JVM won't need to create a new object for the lambda) as well as small functions that are called very often; basically, anything you might use a macro for in C++, you probably want to use aninline fun
orinline val
fortailrec fun
– tail recursion optimizationrun
block function – great way to write code that needs to shortcut (e.g.return@run "NO"
) without having to write a new function and pass every relevant argumentfunctions in functions – functions can be defined within e.g. the
main
function, so again, no having to pass lots of arguments or global variables. Keep in mind that these are represented as objects during runtime. It's too bad they can't beinline
as of yet
Potential pitfalls
Generic wrappers for JVM primitive types can cause TLE for some problems. Use primitive arrays (
IntArray
etc.) whenever possible to avoid this, but see next pointInherits the hack-prone quicksort from Java for primitive arrays. Easiest solution is to use generic arrays or lists instead, but due to the performance benefit of primitive arrays, I've took the trouble to write code that shuffles them randomly before sorting.
For Kotlin versions < 1.3.30, there is a bug that will throw an exception when using
.asJavaRandom()
on instances ofkotlin.random.Random
, including Kotlin's default instance. Either use Java's own Random class, or steal this wrapper:
class JavaRandom(val kt: Random) : java.util.Random(0) {
override fun next(bits: Int): Int = kt.nextBits(bits)
override fun setSeed(seed: Long) {}
}
What does "Pokémon speak" refer to?
Employee employee = new Employee(28, "Jane Doe")
Auto comment: topic has been updated by Spheniscine (previous revision, new revision, compare).
Added new pitfall due to buggy asJavaRandom() function in STL
Wonderful post! I was wondering if there's a way compare two different IntArray or ArrayList as we can do in Python. For eg:
It doesn't look like there are standard library functions to compare lists lexicographically, unfortunately. On the plus side though, you can actually implement the extension function
operator fun <T: Comparable<T>> List<T>.compareTo(other: List<T>)
and thereby use the comparison operators. (Unfortunately though, since there's currently no way to makeList
implementComparable
, it can only go one level. It's one of those tradeoffs of having a type system)Sounds good!
any sample usage/application of your template in a CodeForces problem — http://codeforces.net/contest/1298, perhaps?
I have changed my current template slightly; instead of the "output" second-order function I just define a
PrintWriter.solve()
extension function that's called from the main function. (This is mostly as a partial workaround to this issue with debugging nested inline functions)Examples:
https://codeforces.net/contest/1298/submission/71523046 (practice problem A) https://codeforces.net/contest/1298/submission/71523119 (practice problem B) https://codeforces.net/contest/1298/submission/71523180 (practice problem C)
If you or anyone else wishes to use the template, I'd suggest to at least change the
@author
name, and to replace the0x594E215C123
seed multiplier with any odd number.Still at the very beginning with codeforces, so maybe I'm not seeing the obvious.
How do you set up your environment to use the testset data provided by codeforces?
You can use https://www.jdoodle.com/compile-kotlin-online/ if you are too lazy to setup environment or use https://kotlinlang.org/docs/tutorials/getting-started.html
What is this?
Why do we need to convert Kotlin Random to Java one?
The default sorting algorithm for primitive arrays is quicksort with deterministic pivot selection, as such an adversarial dataset can be constructed to cause it to run in $$$O(n^2)$$$ time.
The Java Random is needed for some rare applications, like when generating a random BigInteger.
Is sorting a real issue or rather an unlikely event? Constructing an adversarial dataset for a dumb quicksort with pivot
a[(i+j)/2]
seems to be a Div1B/C problem, and Java's DualPivotQuicksort seems much much harder to ruin.It is very real if you go through the archives for sorting problems. I haven't learned how to hack it myself, but there are generators around. Many problems in the archives either already have an anti-quicksort dataset in its test cases, or had one added there by a hack.
I was comparing the run time of a kotlin solution with a similar java solution. The Java solution took around 1ms whereas kotlin solution took 300ms. Are you aware of the multiplier used for the kotlin solutions?
Codeforces does not have a multiplier implemented (at this point of time). I've not really tried submitting Java solutions to CF before, but I find that for most problems the time limit is not a big issue, especially if you use the IO techniques I've outlined in this article.
Can we use chelper plugin for kotlin as well ?
I haven't tried it yet
Then can you tell me how do you use the same template for every problem? Do you copy the template every time ?
The basic template (IO and sorting) I leave in for every problem
Other things I keep in a project of code snippets that I copy-paste in
I think I'll give it a try and see if I can work with it; if not I'll stick with what I'm already doing
https://codeforces.net/contest/1436/submission/96587787 got fst by StackOverflow... How to solve it?,
I've not tried to solve that problem yet, but larger graphs do have that problem when using recursive DFS. In fact, large-enough graphs (about $$$|V| \geq 10^6$$$ or so) could cause stack overflow even in C++ programs.
You can either refactor the DFS to be iterative using an explicit stack (ArrayList), or since CF now uses Kotlin 1.4, you can use the DeepRecursiveFunction API, which uses coroutines to simulate a recursive function without using the call stack. I've not tried to use the latter myself however, so I don't know if it's fast enough for this purpose.
Update: My solution with iterative DFS: 96961954. In general this pattern of iterative DFS (either using the bitwise inverse trick I used, or an explicit
visited
array) can solve most graph problems of this type where you need to compute the "sum" of each node's subtree, since there isn't much state needed beyond that which can be put in the result arrays. I might make a blogpost about it if I can get my thoughts organized.Can you share your modint kotlin class ?
97531632
There's quite a bit of code there that I just copy-paste to most problems requiring it. It also includes a
ModIntArray
class backed by anIntArray
.thank you so much, hope you have a good day :)))