I have a dataframe from which I want to draw a random sample--not just any sample but one that contains exactly one randomly sampled row from each of the unique values in the column word:
set.seed(123)
df <- data.frame(
word = sample(LETTERS[1:5], 50, replace = T),
value = sample(1:10, 50, replace = T)
)
head(df)
word value
1 B 1
2 D 5
3 C 8
4 E 2
5 E 6
6 A 3
What I've done to solve this problem is this: 1. Store unique words in vector:
UniqueWords <- unique(df$word)
2. Set up a for loop:
for(i in UniqueWords){
df_sample[i,] <- df[sample(1:nrow(df[df$word==UniqueWords[i], ]), 1), ]
}
The loop, however, does not produce the correct result. How can it be tweaked or, alternatively, what other method can be used?
