MonteCarlo simulation - Create 100 data.frames with rows grouped randomly with a fixed number of subsequent days

Question

Here an example of my data.frame and let's assume that the date col represents days:

df = read.table(text = 'ID date
a 1
a 2
a 3
a 4
a 7
a 12', header = TRUE)

So, my days here range from 1 to 12 and I would like to create 100 data.frames where for each one the date (and ID) col will be grouped randomly by 3 subsequent days.

e.g.

df1

ID date group
    a 1 1   #group 1 = 1, 2, 3
    a 2 1
    a 3 1
    a 4 2   # group 2 = 4, 5, 6
    a 7 3   # group 3 = 7, 8, 9
    a 12 4  # group 4 = 10, 11, 12

df2

ID date group
    a 1 4
    a 2 1  #group 1 = 2, 3, 4
    a 3 1
    a 4 1
    a 7 2  #group 2 = 5, 6, 7  ---  group 3 = 8, 9, 10
    a 12 4 # group 4 = 11, 12 and start again from the beginning 1

df3

ID date group
    a 1 1
    a 2 1
    a 3 2  #group 2 = 3, 4, 5
    a 4 2
    a 7 3  #group 3 = 6, 7, 8 -- group 4 = 9, 10, 11
    a 12 1   #group 1 = 12, 1, 2

etc...

Note that the group col groups the rows by 3 by considering subsequent days which not necessarily appear in the data.frame and the randomness of the whole trick is the start day of group 1.

Do you have any suggestion?

Answer 1

Not sure but you can create an empty list and fill new data frames similar to your example:

set=list()
for(i in 1:100) { set[[i]] = cbind(df,group=sample(rep(c(1,2),each=3))) }