使用返回向量的函数中的行创建一个data.frame

Question

I have data in a number of separate csv files, and I want to create a data.frame with a row for each file. The function below delivers the data to be used for each row. I don't want to change this code, eg include the farmid part of the output vector.

vectorfromfile <- function(farmid) {
    # Reads data from a file named farm{id}.csv, eg 
    # farm001.csv, and returns one named vector
    # of length two with class numeric and names 'apples' 
    # and 'oranges' An example could be c(apples=4, oranges=6)

    # The line below is a dummy for test purposes
    c(apples=farmid+1000, oranges=farmid+2000)
}

I then have a vector, farmids, eg farmids <- c(1,3,5). I need to create a data frame with three columns: id, apples and oranges, and a row for each of the farmids. It should look like the data.frame defined below.

> data.frame(id=c(1,3,5), apples=c(4,2,3), oranges=c(6,5,2) )
  id apples oranges
1  1      4       6
2  3      2       5
3  5      3       2

I have found several ways of doing this, all of them quite ugly and taking up many lines. But I would like to do it in the most elegant way, using the split-apply-combine approach. So I hope I can simply apply to (iterate over) a vector, and get a data.frame as result. Something like

apply(farmids, ???? ) # farmids is  a vector

Is that possible? If not, then perhaps iterating over a list with the same values? And if even that is not possible, what would then be the most elegant way.

My ugly attempts below

vect2df_v1 <- function(farmids=c(1,3,5)) {
    df <- data.frame(id=farmids, apples=rep(NA, length(farmids)), oranges=rep(NA, length(farmids)))
    for (i in 1:length(farmids)) {
       df[i, c('apples', 'oranges')] <- vectorfromfile(df[i, 'id'])
    }
    df
}

vect2df_v2 <- function(farmids=c(1,3,5)) {
    # Obviously it could be written into one (even uglier) line
    farmrow <- function(farmid) { c(farmid, vectorfromfile(farmid)) }
    lst <- lapply(farmids, farmrow)
    mtrx <- matrix(unlist(lst), ncol=3, byrow=T, dimnames=list(NULL,c('id', 'apples','oranges')))
    data.frame(mtrx)
}

Answer 1

This is simple with do.call(rbind, ...) .

You can write your vect2df like this:

vect2df <- function(vec) {
  data.frame(id = vec, do.call(rbind, lapply(vec, vectorfromfile)))
} 

Demo:

vect2df(c(1, 3, 5))
#   id apples oranges
# 1  1   1001    2001
# 2  3   1003    2003
# 3  5   1005    2005

---

Of course, this could all be done pretty directly just using within (if vectorfromfile is not a critical function but can be defined simply.

Example:

within(data.frame(id = c(1, 3, 5)), {
  oranges <- id + 2000
  apples <- id + 1000
})
#   id apples oranges
# 1  1   1001    2001
# 2  3   1003    2003
# 3  5   1005    2005