Random number generators, reproducibility and sampling with dplyr

Let's assume that you want to take some random observations from your data set. Dplyr helps you with the function sample_n(). To make your code reproducible you seed the ID of a “random” set of values. You need to indicate number of rows you want to extract and specify if the rows should be replaced or not. To show you how it works I will use again mtcars dataset which is included in your base R program. Let's see first six rows of this data frame. 

library(dplyr)

data("mtcars")

head(mtcars)

                   mpg cyl disp  hp drat    wt  qsec vs am gear carb

Mazda RX4         21.0   6  160 110 3.90 2.620 16.46  0  1    4    4

Mazda RX4 Wag     21.0   6  160 110 3.90 2.875 17.02  0  1    4    4

Datsun 710        22.8   4  108  93 3.85 2.320 18.61  1  1    4    1

Hornet 4 Drive    21.4   6  258 110 3.08 3.215 19.44  1  0    3    1

Hornet Sportabout 18.7   8  360 175 3.15 3.440 17.02  0  0    3    2

Valiant           18.1   6  225 105 2.76 3.460 20.22  1  0    3    1

sample_n() sample n numbers of random rows 

set.seed(10) 

mtcars%>%sample_n(4,replace = T) # We will take four random rows from mtcars data frame

                   mpg cyl  disp  hp drat    wt  qsec vs am gear carb

Merc 280C           17.8   6 167.6 123 3.92 3.440 18.90  1  0    4    4

Merc 230            22.8   4 140.8  95 3.92 3.150 22.90  1  0    4    2

Merc 280            19.2   6 167.6 123 3.92 3.440 18.30  1  0    4    4

Lincoln Continental 10.4   8 460.0 215 3.00 5.424 17.82  0  0    3    4

sample_frac() is proportional sampling where you need to indicate fraction e.g 0.2 (20%) 

mtcars%>% sample_frac(size=0.2,replace=F) # We will take randomly 20% of mtcars data frame.

                   mpg cyl  disp  hp drat   wt  qsec vs am gear carb

Camaro Z28         13.3   8 350.0 245 3.73 3.84 15.41  0  0    3    4

Cadillac Fleetwood 10.4   8 472.0 205 2.93 5.25 17.98  0  0    3    4

Maserati Bora      15.0   8 301.0 335 3.54 3.57 14.60  0  1    5    8

Merc 280           19.2   6 167.6 123 3.92 3.44 18.30  1  0    4    4

Duster 360         14.3   8 360.0 245 3.21 3.57 15.84  0  0    3    4

Ford Pantera L     15.8   8 351.0 264 4.22 3.17 14.50  0  1    5    4 

In terms of testing for sampling error, in case of big datasets and large sample size, both methods (random and proportional) deliver similar results. Proportional sampling is a better approach for smaller datasets, for smaller sample sizes and if relative group proportions matter.