Applying function using multiple columns as argument , function returns a data.frame

Question

I am trying to apply a function that uses multiple columns of a dataframe as arguments, with the function returning a dataframe for each row. I can use a for loop here, but Wanted to check if there is any other way of doing this

A simple example is being provided here. my original problem is slightly more complicated.

DF1<-data.frame(start=seq(from=1, to=5, by=1),end=seq(from=10, to=14, by=1))

rep_fun <- function(x,y)
{
  data.frame( A=seq(x, y)) #produces a sequence between x and y
}

DF2<-data.frame()
for (i in 1:nrow(DF1)){
  temp<-data.frame(rep_fun(DF1$start[i],DF1$end[i]))
 DF2<-rbind(temp,DF2) # this contains a dataframe that has a sequence between 'start' and 'end' for  each row in DF1 

}

The desired result which I am able to obtain through a for-loop is shown below. Not all rows are being shown here. Rows 1 to 10, shows the sequence corresponding to row 5 in DF1

I trust this is a case of the xy problem. But here is a possible approach: as.vector(apply(DF1, 1, function(x) x[1] : x[2])) — missuse
– missuse, Commented Nov 4, 2017 at 12:52
As I had mentioned, my original problem is more complicated. The rep_fun equivalent in my original problem will return a multi row and multi column data frame. — ashleych
– ashleych, Commented Nov 4, 2017 at 12:57

G. Grothendieck · Accepted Answer · 2017-11-04 17:44:40Z

3

1) lapply Split DF1 by nrow(DF1):1 so that it comes out in reverse order and then lapply over that list and rbind its components together. No packages are used.

DF3 <- do.call("rbind", lapply(split(DF1, nrow(DF1):1), with, rep_fun(start, end)))
rownames(DF3) <- NULL

identical(DF2, DF3)
## [1] TRUE

2) Map or this alternative:

fun <- function(x) with(x, rep_fun(start, end))
DF4 <- do.call("rbind", Map(fun, split(DF1, nrow(DF1):1), USE.NAMES = FALSE))

identical(DF4, DF2)
## [1] TRUE

3) Map/rev Like (2) this uses Map but this time using rep_fun directly. Also, it uses rev to order the output after the computation rather than split to order the input before the computation.

DF5 <- do.call("rbind", with(DF1, rev(Map(rep_fun, start, end))))

identical(DF5, DF2)
## [1] TRUE

edited Nov 4, 2017 at 17:44

answered Nov 4, 2017 at 15:30

G. Grothendieck

273k18 gold badges221 silver badges365 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

ashleych Over a year ago

The reversing thing wasnt a requirement to be honest, but great to find out that is possible! Option 3 is the one I am going with. Its about 3 times faster than the For loop option in my original problem!

Collectives™ on Stack Overflow

Applying function using multiple columns as argument , function returns a data.frame

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related