I have data frame where I want to shuffle or more generally sort the values hierachially but I am stuck. Here is my example:
library(plyr)
set.seed(123)
df <- data.frame(a = rep(letters[1:4], each = 10), b = rnorm(40))
> head(df)
a b
1 a -0.1264280
2 a 0.7284234
3 a -1.8782385
4 a 0.2530623
5 a 0.7577013
6 a -0.9339964
In this example, I want to suffle (sample) the values but only within the letters so that a value assigned to the letter a in colum a cannot be assigned to any other letter but only to a different a row but not b, c or d.
I've tried this ddply(df, c('a'), b = sample(b)) but this didn't work.
Using dplyr, group_by(a) then use mutate(b = sample(b))
library(dplyr)
head(df, 10)
a b
1 a -0.56047565
2 a -0.23017749
3 a 1.55870831
4 a 0.07050839
5 a 0.12928774
6 a 1.71506499
7 a 0.46091621
8 a -1.26506123
9 a -0.68685285
10 a -0.44566197
df %>% group_by(a) %>% mutate(b = sample(b))
# A tibble: 40 x 2
# Groups: a [4]
a b
<chr> <dbl>
1 a 1.56
2 a 0.461
3 a 0.0705
4 a 1.72
5 a -0.560
6 a -0.446
7 a -1.27
8 a 0.129
9 a -0.230
10 a -0.687
# ... with 30 more rows