Replicate(): A Robust alternative to traditional ‘for loop’ iterations

Introduction

Consider the amount of time that “for loops” use to run a function or expression a number of times, say 10,000 times. Understandably, it will take a lot of time, but the Replicate function offers a simpler and faster solution.

Replicate() makes it possible to produce repeated outcomes which can be used for generating test data sets, statistical analysis, simulations, and other uses.

Basic syntax:

replicate(n, expr)

Where,
n: the number of times to replicate an expression or function
expr: is a statement or function to evaluate

Example 1: Replicate an expression Multiple Times

#Printing "Sankhya World" 4 times
print(replicate(4, "Sankhya World"))

[1] "Sankhya World" "Sankhya World" "Sankhya World" "Sankhya World"

Try it yourself: replace 4 by 500 and also, run the below code.
for (i in 1:500) { print(“Sankhya World”)}

Observe the time taken for the same task using “replicate()” and “for loop”. Even while it might only seem like a few milliseconds, we can only think how much time will be saved when dealing with massive volumes of data and thousands of loop runs.

Example 2: Probability of Toss

Toss<- c("Heads","Tails")
Toss_results <- replicate(25, sample(Toss, 1, replace = TRUE))
Toss_results

 [1] "Heads" "Tails" "Tails" "Heads" "Heads" "Tails" "Tails" "Heads" "Tails"
[10] "Tails" "Tails" "Heads" "Tails" "Heads" "Tails" "Heads" "Tails" "Heads"
[19] "Tails" "Heads" "Tails" "Heads" "Heads" "Heads" "Heads"

table(Toss_results)

Toss_results
Heads Tails 
   13    12 

Here, sample(Toss, 1, replace = TRUE) randomly selects one value either “Heads” or “Tails”, simulating a single coin flip. replicate(25, …) repeats this simulation 25 times, giving you a vector of 25 Toss results.

Example 3: Replicating a function multiple times

Creating sample data sets which can be used as an input to model or testing algorithms

# Calculating mean and sd of 1 to 10 numbers
mean_val<-mean(1:10)
sd_val<-sd(1:10)

# replicate normal distribution with mean and sd
rnormd_df<-replicate(4, rnorm(3, mean_val,sd_val))
rnormd_df

         [,1]      [,2]      [,3]      [,4]
[1,] 4.019644 3.8421126  7.331920 7.4974377
[2,] 2.169128 3.4926379 10.222299 0.1086222
[3,] 2.052114 0.6676298  9.312386 1.0337011

#Getting summary statistics for each sample(column)
summary(rnormd_df)

       V1              V2               V3               V4        
 Min.   :2.052   Min.   :0.6676   Min.   : 7.332   Min.   :0.1086  
 1st Qu.:2.111   1st Qu.:2.0801   1st Qu.: 8.322   1st Qu.:0.5712  
 Median :2.169   Median :3.4926   Median : 9.312   Median :1.0337  
 Mean   :2.747   Mean   :2.6675   Mean   : 8.956   Mean   :2.8799  
 3rd Qu.:3.094   3rd Qu.:3.6674   3rd Qu.: 9.767   3rd Qu.:4.2656  
 Max.   :4.020   Max.   :3.8421   Max.   :10.222   Max.   :7.4974  

Here, rnorm(3, mean_val,sd_val) creates our first sample with 3 set of values for the the mean and sd of 1 to 10 numbers. replicate(6,…) further repeats it for 4 times giving us a matrix of [3,4] i.e. 4 samples.

Conclusion:

This time-saving function can greatly reduce the complexity of your repetitive coding jobs and thus optimize the code.