add rows to group in df

Question

I have a data.frame with varying numbers of rows per person. If a person has less than 4 rows I would like to give them 4 rows. I would like the extra rows to be a replicate of the the last row per person e.g., Tom below.

I know how to repeat rows by using this code:

dff <- df[rep(1:nrow(df), each=4), ] 

But the problem here is that I have no argument for only replicating rows for IDs which occur less than 4 times.

I've tried this:

dff <- ifelse(length(df$noms < 4), 
              df[rep(1:nrow(df), each=4), ], df) ##doesn't work

I think I could use do.call(rbind, ...) but again, I can't manage to integrate my argument into the code.

Here is something I've tried:

df <- ifelse(length(df$noms<4), 
             do.call(rbind, by(df, df$noms, rbind, 1:nrow(df))), 
             df) ## doesn't work

And also I have not managed to include an argument for how many repeats to do.

   noms fruits apple orange kiwi all_comb comb
1  mary  apple     1      0    0        1    1
2  mary  grape     0      0    0        0    1
3  mary orange     0      1    0        0    1
4  mary  apple     1      0    0        1    1
5  john banana     0      0    0        0    1
6  john  apple     1      0    0        1    1
7  john  apple     1      0    0        1    1
8  john  apple     1      0    0        1    1
9  lucy   kiwi     0      0    1        0    1
10 lucy orange     0      1    0        0    1
11 lucy  apple     1      0    0        1    1
12 lucy  berry     0      0    0        0    1
13  tom orange     0      1    0        0    1

Desired output

   noms fruits apple orange kiwi all_comb comb
1  mary  apple     1      0    0        1    1
2  mary  grape     0      0    0        0    1
3  mary orange     0      1    0        0    1
4  mary  apple     1      0    0        1    1
5  john banana     0      0    0        0    1
6  john  apple     1      0    0        1    1
7  john  apple     1      0    0        1    1
8  john  apple     1      0    0        1    1
9  lucy   kiwi     0      0    1        0    1
10 lucy orange     0      1    0        0    1
11 lucy  apple     1      0    0        1    1
12 lucy  berry     0      0    0        0    1
13  tom orange     0      1    0        0    1
14  tom orange     0      1    0        0    1
15  tom orange     0      1    0        0    1
16  tom orange     0      1    0        0    1

Here are some reproducible data:

noms <- as.character(c('mary', 'mary','mary','mary','john','john','john',
                       'john','lucy','lucy','lucy','lucy', 'tom'))
fruits <- as.character(c('apple','grape','orange','apple','banana',
                         'apple','apple','apple','kiwi','orange',
                         'apple','berry', 'orange'))
df <- data.frame(noms,fruits)

Answer 1

You're on the right track with ifelse. Try the following:

noms <- as.character(c('mary', 'mary','mary','mary','john','john','john',
                       'john','lucy','lucy','lucy','lucy', 'tom'))
fruits <- as.character(c('apple','grape','orange','apple','banana',
                         'apple','apple','apple','kiwi','orange',
                         'apple','berry', 'orange'))
df <- data.frame(noms,fruits)

x <- with(df, ave(rep(1, nrow(df)), noms, FUN = length))
df[rep(rownames(df), ifelse(x >= 4, 1, 4)), ]
#      noms fruits
# 1    mary  apple
# 2    mary  grape
# 3    mary orange
# 4    mary  apple
# 5    john banana
# 6    john  apple
# 7    john  apple
# 8    john  apple
# 9    lucy   kiwi
# 10   lucy orange
# 11   lucy  apple
# 12   lucy  berry
# 13    tom orange
# 13.1  tom orange
# 13.2  tom orange
# 13.3  tom orange

Answer 2

using code from a previous question; ddply to split and add rows to each group

here is the answer

first, create a variable called numbers in df

df$numbers<-ave(df$noms,df$noms, FUN=seq_along)

then manipulate code given in previous answer:

AddRows <- function(df) {
  new_numbers <- seq(from = min(df$numbers), to = 4)
  new_numbers <- new_numbers[new_numbers != 0]    ##probably don't need this line
  noms <- rep(unique(df$noms), length(new_numbers))

  return(data.frame(df, new_numbers))
}

ddply(df, .(noms), AddRows)