Combine data frames by rows in respect to even odd elements

Question

I have two data frames:

> df1 <- data.frame(a=3:4, b=1:2)
> df2 <- data.frame(a=5:6, b=7:8)
>
> df1
  a b
1 3 1
2 4 2
>
> df2
  a b
1 5 7
2 6 8

How can I combine (or zip or interleave) them by rows so the result will be following:

> df_final <- some_magic_function(df1, df2)
> df_final
1 3 1
2 5 7
3 4 2
4 6 8

I've found similar approaches here and there, but it is for vectors:

> a <- c(1, 2, 3)
> b <- c(4, 5, 6)
>
> a
[1] 1 2 3
> b
[1] 4 5 6
>
> as.vector(rbind(a,b))
[1] 1 4 2 5 3 6
>
> z1 <- setNames(c(a,b),rep(seq_along(a),2))
> z1 <- as.vector(z1[order(names(z1))]) 
> z1
[1] 1 4 2 5 3 6
>
> c( matrix(c(a,b), nrow=2, byrow=TRUE) ) 
[1] 1 4 2 5 3 6
>
> c(a,b)[ order( c(seq_along(a), seq_along(b)))]

How to do the similar for data frames? Please note I want to interleave rows, not to concatenate data frames and then somehow rearrange it's rows.

tushaR · Answer 1 · 2017-10-17T12:22:22.597

2

Try this:

as.data.frame.matrix(mapply(function(x,y){rbind(x,y)},df1,df2))

#  a b
#1 3 1
#2 5 7
#3 4 2
#4 6 8

or

Reduce(function(x,y){rbind(x,y)},(sapply(1:nrow(df1),
 functio‌n(i,x,y){rbind(x[i,]‌,y[i,])},df1,df2,sim‌plify = FALSE)))

edited Oct 17 '17 at 12:22

answered Oct 17 '17 at 11:13

tushaR

3,083
1
20
33

while the OP example is all numeric, preserving column types is likely desirable. – hrbrmstr Oct 17 '17 at 11:17
@hrbrmstr Sorry, but I did not get the point you are trying to make. Could you please explain in detail the issues my approach would lead to? – tushaR Oct 17 '17 at 11:22
1

your version turns it into a matrix along the way. by default that means column types get converted. if the OP's real data frame has a mix of types, this will coerce them into the same type. – hrbrmstr Oct 17 '17 at 11:27
@hrbrmstr Got it. Thanks a lot. – tushaR Oct 17 '17 at 11:28
3

For data frames with cols of the same type, this is definitely a clever solution! – hrbrmstr Oct 17 '17 at 11:29
1

@hrbrmstr: How about this approach: `Reduce(function(x,y){rbind(x,y)},(sapply(1:nrow(df1),function(i,x,y){rbind(x[i,],y[i,])},df1,df2,simplify = F)))` ? – tushaR Oct 17 '17 at 12:08

hrbrmstr · Accepted Answer · 2017-10-17T11:24:59.327

These preserves column types but 2 of them rely on an internal ggplot2 function.

The easy way:

gdata::interleave(df1, df2)

The hard way(s):

Base:

do.call(
  cbind.data.frame, 
  setNames(
    lapply(
      colnames(df1), 
        function(x) { 
          ggplot2:::interleave(df1[,x], df2[,x])
        }
    ), colnames(df1)
  )
)

Tidyverse:

library(tidyverse)

map(colnames(df1), ~ggplot2:::interleave(df1[,.x], df2[,.x])) %>% 
  set_names(colnames(df1)) %>% 
  bind_cols()

score 1 · Answer 3 · answered Oct 17 '17 at 11:26

1

A slightly over-fitted, Vectorized solution could be,

full_d <- rbind(df1, df2)
rbind(full_d[c(TRUE, FALSE),], full_d[c(FALSE, TRUE),])

answered Oct 17 '17 at 11:26

Sotos

51,121
6
32
66

Combine data frames by rows in respect to even odd elements

3 Answers3