Split and unsplit a dataframe in four parts - r

Question

I'd like to split a dataframe in 4 equals parts, because I'd like to use the 4 cores of my computer.
I did this :
df2 <- split(df, 1:4)
unsplit(df2, f=1:4)
and that
df2 <- split(df, 1:4)
unsplit(df2, f=c('1','2','3','4')
But the unsplit function did not work, I have these warnings messages
1: In split.default(seq_along(x), f, drop = drop, ...) :
data length is not a multiple of split variable
...
Do you have an idea of the reason ?

Answer 1

How many rows in df? You will get that warning if the number of rows in your table is not divisible by 4. I think you are using the split factor f incorrectly, unless what you want to do is put each subsequent row into a different split data.frame.
If you really want to split your data into 4 dataframes. one row after the other then make your splitting factor the same size as the number of rows in your dataframe using rep_len like this:
## Split like this:
split(df , f = rep_len(1:4, nrow(df) ) )
## Unsplit like this:
unsplit( split(df , f = rep_len(1:4, nrow(df) ) ) , f = rep_len(1:4,nrow(df) ) )
Hopefully this example illustrates why the error occurs and how to avoid it (i.e. use a proper splitting factor!).
## Want to split our data.frame into two halves, but rows not divisible by 2
df <- data.frame( x = runif(5) )
df
## Splitting still works but...
## We get a warning because the split factor 'f' was not recycled as a multiple of it's length
split( df , f = 1:2 )
#$`1`
# x
#1 0.6970968
#3 0.5614762
#5 0.5910995
#$`2`
# x
#2 0.6206521
#4 0.1798006
Warning message:
In split.default(x = seq_len(nrow(x)), f = f, drop = drop, ...) :
data length is not a multiple of split variable
## Instead let's use the same split levels (1:2)...
## but make it equal to the length of the rows in the table:
splt <- rep_len( 1:2 , nrow(df) )
splt
#[1] 1 2 1 2 1
## Split works, and f is not recycled because there are
## the same number of values in 'f' as rows in the table
split( df , f = splt )
#$`1`
# x
#1 0.6970968
#3 0.5614762
#5 0.5910995
#$`2`
# x
#2 0.6206521
#4 0.1798006
## And unsplitting then works as expected and reconstructs our original data.frame
unsplit( split( df , f = splt ) , f = splt )
# x
#1 0.6970968
#2 0.6206521
#3 0.5614762
#4 0.1798006
#5 0.5910995

Answer 2

In the R language 'split' example . . .
aq <- airquality
g <- aq$Month
l <- split(aq,g)
After the 'scale' function is executed
l <- lapply(l, transform, Ozone = scale(Ozone))
I am guessing that at one time in R history
the function 'scale' did not add extra attributes
to the column it is modifying.
..$ Ozone : num ...
.. ..- attr(*, "scaled:center")= num 29.4
.. ..- attr(*, "scaled:scale")= num 18.2
As seen in here . . .
> str(l)
List of 5
$ 5:'data.frame': 31 obs. of 6 variables:
..$ Ozone : num [1:31, 1] 0.782 0.557 -0.523 -0.253 NA ...
.. ..- attr(*, "scaled:center")= num 23.6
.. ..- attr(*, "scaled:scale")= num 22.2
..$ Solar.R: int [1:31] 190 118 149 313 NA NA 299 99 19 194 ...
..$ Wind : num [1:31] 7.4 8 12.6 11.5 14.3 14.9 8.6 13.8 20.1 8.6 ...
..$ Temp : int [1:31] 67 72 74 62 56 66 65 59 61 69 ...
..$ Month : int [1:31] 5 5 5 5 5 5 5 5 5 5 ...
..$ Day : int [1:31] 1 2 3 4 5 6 7 8 9 10 ...
$ 6:'data.frame': 30 obs. of 6 variables:
..$ Ozone : num [1:30, 1] NA NA NA NA NA ...
.. ..- attr(*, "scaled:center")= num 29.4
.. ..- attr(*, "scaled:scale")= num 18.2
..$ Solar.R: int [1:30] 286 287 242 186 220 264 127 273 291 323 ...
..$ Wind : num [1:30] 8.6 9.7 16.1 9.2 8.6 14.3 9.7 6.9 13.8 11.5 ...
..$ Temp : int [1:30] 78 74 67 84 85 79 82 87 90 87 ...
..$ Month : int [1:30] 6 6 6 6 6 6 6 6 6 6 ...
..$ Day : int [1:30] 1 2 3 4 5 6 7 8 9 10 ...
$ 7:'data.frame': 31 obs. of 6 variables:
..$ Ozone : num [1:31, 1] 2.399 -0.32 -0.857 NA 0.154 ...
.. ..- attr(*, "scaled:center")= num 59.1
.. ..- attr(*, "scaled:scale")= num 31.6
..$ Solar.R: int [1:31] 269 248 236 101 175 314 276 267 272 175 ...
..$ Wind : num [1:31] 4.1 9.2 9.2 10.9 4.6 10.9 5.1 6.3 5.7 7.4 ...
..$ Temp : int [1:31] 84 85 81 84 83 83 88 92 92 89 ...
..$ Month : int [1:31] 7 7 7 7 7 7 7 7 7 7 ...
..$ Day : int [1:31] 1 2 3 4 5 6 7 8 9 10 ...
$ 8:'data.frame': 31 obs. of 6 variables:
..$ Ozone : num [1:31, 1] -0.528 -1.284 -1.108 0.455 -0.629 ...
.. ..- attr(*, "scaled:center")= num 60
.. ..- attr(*, "scaled:scale")= num 39.7
..$ Solar.R: int [1:31] 83 24 77 NA NA NA 255 229 207 222 ...
..$ Wind : num [1:31] 6.9 13.8 7.4 6.9 7.4 4.6 4 10.3 8 8.6 ...
..$ Temp : int [1:31] 81 81 82 86 85 87 89 90 90 92 ...
..$ Month : int [1:31] 8 8 8 8 8 8 8 8 8 8 ...
..$ Day : int [1:31] 1 2 3 4 5 6 7 8 9 10 ...
$ 9:'data.frame': 30 obs. of 6 variables:
..$ Ozone : num [1:30, 1] 2.674 1.928 1.721 2.467 0.644 ...
.. ..- attr(*, "scaled:center")= num 31.4
.. ..- attr(*, "scaled:scale")= num 24.1
..$ Solar.R: int [1:30] 167 197 183 189 95 92 252 220 230 259 ...
..$ Wind : num [1:30] 6.9 5.1 2.8 4.6 7.4 15.5 10.9 10.3 10.9 9.7 ...
..$ Temp : int [1:30] 91 92 93 93 87 84 80 78 75 73 ...
..$ Month : int [1:30] 9 9 9 9 9 9 9 9 9 9 ...
..$ Day : int [1:30] 1 2 3 4 5 6 7 8 9 10 ...
But now it does add those attributes
..$ Ozone : num ...
.. ..- attr(*, "scaled:center")= num 29.4
.. ..- attr(*, "scaled:scale")= num 18.2
and the very simple 'unsplit' function is not programmed to handle those attributes.
> unsplit(l,g)
Error in xj[i, , drop = FALSE] : (subscript) logical subscript too long
The (direct and simple) solution is to get rid of those attributes.
attributes(l[[1]]$Ozone) <- NULL
attributes(l[[2]]$Ozone) <- NULL
attributes(l[[3]]$Ozone) <- NULL
attributes(l[[4]]$Ozone) <- NULL
attributes(l[[5]]$Ozone) <- NULL
Then try to unsplit again.
str( unsplit(l,g) )
> str( unsplit(l,g) )
'data.frame': 153 obs. of 6 variables:
$ Ozone : num 0.782 0.557 -0.523 -0.253 NA ...
$ Solar.R: int 190 118 149 313 NA NA 299 99 19 194 ...
$ Wind : num 7.4 8 12.6 11.5 14.3 14.9 8.6 13.8 20.1 8.6 ...
$ Temp : int 67 72 74 62 56 66 65 59 61 69 ...
$ Month : int 5 5 5 5 5 5 5 5 5 5 ...
$ Day : int 1 2 3 4 5 6 7 8 9 10 ...
So, now it works.
Andre Mikulec