Unsplit a list of data frames after subsetting the data frames in the list

Question

Let's say I have a data frame "x"

> x
    x1         x2 x3
1  box 0.81432465  4
2  box 0.19628122  2
3  box 0.06619734  1
4  box 0.90403568  5
5  box 0.52693274  3
6  axe 0.28665840  2
7  axe 0.45193228  3
8  axe 0.48278466  4
9  axe 0.86553847  5
10 axe 0.13925190  1
11 cat 0.86340413  5
12 cat 0.28387540  2
13 cat 0.24297445  1
14 cat 0.36651366  3
15 cat 0.55038108  4

Then I perform following operations on it

> x.factor <- factor(x[,1]) ## convert column 1 as factors
> x.split <- split(x, x.factor)
> unsplit(x.split, x.factor) ## get back original data frame

works fine till now. But when I do this, it gives me an error

> x.split2 <- lapply(x.split, function(x) {head(x,1)})
> unsplit(x.split2, x.factor) ## trying to combine into a data frame

Error in `row.names<-.data.frame`(`*tmp*`, value = value) : 
  duplicate 'row.names' are not allowed
In addition: Warning message:
non-unique values when setting 'row.names': ‘1’, ‘11’, ‘6’

I don't get it because, if I print out x.split2, the row names are unique for each element in the list.

Why am I getting this error?

Answer 1

Instead of unsplit, you can use the common do.call(rbind, ...) approach:

do.call(rbind, x.split2)
#      x1        x2 x3
# axe axe 0.2866584  2
# box box 0.8143246  4
# cat cat 0.8634041  5

Your present approach doesn't work because your "x.factor" object has duplicated values that relate to the original number of rows in your data.frame. Since you're just taking one of each factor, you can also try something like the following:

unsplit(x.split2, levels(x.factor))
#     x1        x2 x3
# 6  axe 0.2866584  2
# 1  box 0.8143246  4
# 11 cat 0.8634041  5