custom grouped dplyr function (sample_n)

Question

I am trying to apply a sampling function in a grouped fashion to a data frame, where it should sample n samples from each group, or all group members if the group size is smaller than n.

Using dplyr, I first tried

library(dplyr)
mtcars %>% group_by(cyl) %>% sample_n(2)

This works when n is smaller than all the group sizes but does not take the full group when I choose n larger than the group size (note that there are 7 cars in one of the cyl groups):

mtcars %>% group_by(cyl) %>% sample_n(8)
Error: `size` must be less or equal than 7 (size of data), 
set `replace` = TRUE to use sampling with replacement

I tried to solve this by creating an adapted group_n function like so:

sample_n_or_all <- function(tbl, n) {
  if (nrow(tbl) < n)return(tbl)
  sample_n(tbl, n)
}

but using my custom function (mtcars %>% group_by(cyl) %>% sample_n_or_all(8)) generates the same error.

Any suggestions how I can adapt my function so I can apply it to each of the groups? Or another solution to the problem?

Answer 1

We could check the number of rows in the group and pass the value to sample_n accordingly.

library(dplyr)
n <- 8

temp <- mtcars %>% group_by(cyl) %>% sample_n(if(n() < n) n() else n) 
temp

#    mpg   cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb
#   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1  21.4     4 121     109  4.11  2.78  18.6     1     1     4     2
# 2  27.3     4  79      66  4.08  1.94  18.9     1     1     4     1
# 3  24.4     4 147.     62  3.69  3.19  20       1     0     4     2
# 4  22.8     4 108      93  3.85  2.32  18.6     1     1     4     1
# 5  26       4 120.     91  4.43  2.14  16.7     0     1     5     2
# 6  33.9     4  71.1    65  4.22  1.84  19.9     1     1     4     1
# 7  30.4     4  75.7    52  4.93  1.62  18.5     1     1     4     2
# 8  30.4     4  95.1   113  3.77  1.51  16.9     1     1     5     2
# 9  21       6 160     110  3.9   2.62  16.5     0     1     4     4
#10  17.8     6 168.    123  3.92  3.44  18.9     1     0     4     4
# … with 13 more rows

We can check number of rows in each group after that.

table(temp$cyl)

#4 6 8 
#8 7 8 

table(mtcars$cyl)

# 4  6  8 
#11  7 14 

Answer 2

We can do this without using a logical condition with pmin

library(dplyr)
tmp <- mtcars %>%
         group_by(cyl) %>%
         sample_n(pmin(n(), n))
# A tibble: 23 x 11
# Groups:   cyl [3]
#     mpg   cyl  disp    hp  drat    wt  qsec    vs    am  gear  carb
#   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1  33.9     4  71.1    65  4.22  1.84  19.9     1     1     4     1
# 2  27.3     4  79      66  4.08  1.94  18.9     1     1     4     1
# 3  21.4     4 121     109  4.11  2.78  18.6     1     1     4     2
# 4  30.4     4  75.7    52  4.93  1.62  18.5     1     1     4     2
# 5  21.5     4 120.     97  3.7   2.46  20.0     1     0     3     1
# 6  32.4     4  78.7    66  4.08  2.2   19.5     1     1     4     1
# 7  30.4     4  95.1   113  3.77  1.51  16.9     1     1     5     2
# 8  26       4 120.     91  4.43  2.14  16.7     0     1     5     2
# 9  17.8     6 168.    123  3.92  3.44  18.9     1     0     4     4
#10  21       6 160     110  3.9   2.62  16.5     0     1     4     4
# … with 13 more rows

-checking

table(tmp$cyl)
# 4 6 8 
# 8 7 8