Skip to main content

Process text file concurrently

 Process the same input file concurrently Let's say we have a text file that contains 2 different types of lines.  For the sake of example, one group starts with "i:" (for Integer) and the other group starts with "s:" (for String), something like the following file: s:New york s:Apple i:387548 s:Amsterdam i:4556 i:39874 s:Orange i:56787 s:Banana i:4657567 s:Turkey i:45679456 s:Iran i:4356456 i:23423 i:456 s:Ukraine i:453645 i:5456 We want to process these input lines separately but concurrently. And the process can be anything, for this example, we just print them to the console. But what matters is that we open the file only for READ so 2 different processes (or routines) can read the file at the same time. The Cats Effect answer to open files (or in general any resources) in a safe and efficient way, is  Resource class. So we need a function to take the path to the file and give us a Resource of some type that allows us to read the file line by line. One o...

Using IO class to perform side-effects in a pure functional way

 Cats-effects' IO class is the answer of the library to performing side effects (or simply effects) in a functional way. Functional programming is focused on pure functions and a pure function is a function that does not perform any side effects but only produces an output based on the receiving inputs. Real-world programs are far from this mathematical abstract definition. A program with no side effects is doing nothing as most of the useful things are side effects, such as writing to a database, writing to a file, reading from a file, printing on the screen, and ...



An IO class is actually a description of a computation, rather than the execution of that computation. That's why it will not make a function an impure function.

Example:

Printing on the screen (which is a side effect), as some sort of logging.

If we want to print on the screen using inside the body of a function the side effect happens but if we return the effect as an IO object wrapping the calculation result, then we have a pure function that delegates the effects to its caller.


Bad function (not pure function):

  def notPureMultiple(a: Int, b: Int): Int = {
    // some kind of logging inside the function
    println(s"not pure: calculating $a * $b")
    a * b
  }

Good function (pure function):
  def pureMultiple(a: Int, b: Int): IO[Int] = {
    IO {
      println(s"pure: calculating $a * $b")
      a * b
    }
  }


Complete code:

package com.blogspot.scaland7.catseffects

import cats.effect.IO
import cats.effect.unsafe.implicits.global

object SideEffects01 {

  def notPureMultiple(a: Int, b: Int): Int = {
    // some kind of logging inside the function
    println(s"not pure: calculating $a * $b")
    a * b
  }

  def pureMultiple(a: Int, b: Int): IO[Int] = {
    IO {
      println(s"pure: calculating $a * $b")
      a * b
    }
  }

  def main(args: Array[String]): Unit = {
    // this calculates the result and do the side effect right away, which is not good!
    val value1: Int = notPureMultiple(10, 20)

    // this return an IO of the result so the side effect has not happened until we run the io
    val value2: IO[Int] = pureMultiple(20, 10)
    value2.unsafeRunSync() // effects happens here
  }
}

For cats-effects dependency add the following line to your build.sbt file

libraryDependencies += "org.typelevel" %% "cats-effect" % "3.4.8"

Comments

Popular posts from this blog

Functional algorithm to find the position of a sub-string in a string

Let's practice some functional problem-solving!  Consider the classic problem of finding the index of a sub-string in a larger string. So we want to write a function that if we pass it the following strings it returns 0 . "hello world", "hello" If we pass it the following strings it returns 6 "hello world", "world" If we pass it the following strings it returns None , because "apple" can not be found inside "hello world. "hello world", "apple" So our function signature should look like this: def findFirstSubString(str: String, subStr: String): Option[Int] = ??? If we want to implement it in an imperative way, and yes Scala allows imperative programming, we can implement the function as follows, but we don't want imperative programming, do we? def findFirstSubString_bad(str: String, subStr: String): Option[Int] = { var j = 0; var temp = "" for (i So let's find the functional algorith...

Find a pair in an array for the given sum

Algorithm: Find a pair in an array of numbers for a given sum Given an Array (or List) and a sum, we want to find a pair of numbers whose sum will be equal to the given sum.  For example for the [1,4,7] and 5 the outcome should be pair of 1 and 4. Or for [5, 7, 2, 8, 3] and 15 the outcome should be pair of 7 and 8.  Both pairs of 1 and 5, and 2 and 4 are correct for the input of [2, 4, 1, 5] and 6. The function signature looks like this: def findPairForSum(list: List[Int], sum: Int): Option[(Int, Int)] = ??? The return type is  Option[(Int, Int)] because there might be no pair at all. There are two approaches to solving this problem: The brute force approach which is the simplest and most naive one The first sort approach which is more efficient and a bit more complex Naive approach We can check the given sum against all possible pair combinations. In imperative programming, it would be the famous i and j for loops. But in functional programming, we can implement this us...

Functional algorithm to find how many times a sub-string occurs in a string

Another String search problem that we like to solve in a functional way. This problem is very similar to the previous one,  Functional algorithm to find the position of a sub-string in a string , but it is a bit easier in the sense that we can always return an Int , so we do not have to return an Option of Int , because the number of the occurrences of a String in another String is always a non-negative number. So the signature of the function that we want to write looks like this: def countSubstring(str: String, subString: String): Int = ??? ⚠️ Be careful about overthinking the problems and always remain within the scope, in this case, it is very easy to overthink this problem by mixing up the terms sub-string and word . We only care about sub-strings and our algorithm does not know about words, meaning that our function should return the number 3 ( and not 2 )  if we pass it the following parameter: "This book is the best book among my books", "book" As we saw...