CODEWITHSUNDEEP
rdataframe
d12n
d12n

Reputation: 861

Copy pasting column values in a dataframe

I have a dataframe that looks something like this:

TRUE    x_value1    x_value2    y_value1    y_value2    x_id    y_id
0           1            4        11             14      1       7
1           2            5        12             15      2       8
0           3            6        13             16      3       9

And I would like to add rows to this data frame and switch x and y, so that in row 4 x_value1=y_value1 and x_id=y_id and y_value1=x_value1...etc.

like this:

 TRUE   x_value1    x_value2    y_value1    y_value2    x_id    y_id
    0           1            4        11             14      1       7
    1           2            5        12             15      2       8
    0           3            6        13             16      3       9
    0           11           14       1              4       7       1
    1           12           15       2              5       8       2
    0           13           16       3              6       9       3

I can do this with for loops, but that takes ages

e.g.

for (i in 1:3)
{ dataframe[i+3,2]<-dataframe[i,4] //for i=1, finds 4th row, column "x_value1" and switches it with first row, column "y_value1"
  dataframe[i+3,3]<-dataframe[i,5]
...etc.
}

The example data frame I have (the first table above):

structure(list(TRUE. = c(0L, 1L, 0L), x_value1 = 1:3, x_value2 = 4:6, 
    y_value1 = 11:13, y_value2 = 14:16, x_id = 1:3, y_id = 7:9), .Names = c("TRUE.", 
"x_value1", "x_value2", "y_value1", "y_value2", "x_id", "y_id"
), row.names = c(NA, 3L), class = "data.frame")

Desired (second table above):

structure(list(TRUE. = c(0L, 1L, 0L), x_value1 = 1:3, x_value2 = 4:6, 
    y_value1 = 11:13, y_value2 = 14:16, x_id = 1:3, y_id = 7:9), .Names = c("TRUE.", 
"x_value1", "x_value2", "y_value1", "y_value2", "x_id", "y_id"
), row.names = c(NA, 3L), class = "data.frame")

Upvotes: 2

Views: 5964

Answers (3)

janattack
janattack

Reputation: 420

Instead of rearranging the data frames as others have suggested, you could just rename columns and then let rbind take care of matching up the order:

dat2 <- dat
colnames (dat2) <- colnames (dat) [c (1, 4, 5, 2, 3, 7, 6)]

dat3 <- rbind (dat, dat2)

EDIT: Or more programmatically, to avoid having to type in column indexes, which I keep doing wrong

dat2 <- dat

newnames <- gsub ("(x|y)_", "\\1\\1_", colnames(dat))
newnames <- gsub ("xx_", "y_", newnames)
newnames <- gsub ("yy_", "x_", newnames)
colnames(dat2) <- newnames

dat3 <- rbind (dat, dat2)

Upvotes: 2

IRTFM
IRTFM

Reputation: 263342

To do it more "programatically": use grep with a pattern for the column names:

> grep("y_", names(dat))
[1] 4 5 7
> grep("x_", names(dat))
[1] 2 3 6

dat2 <- dat
# Replace all the dat2 "y_"-names with "x_"-names
colnames(dat2)[grep("y_", colnames(dat))] <- 
                    colnames(dat)[grep("x_", colnames(dat))]
# Replace alldat2  the "x_"-names with "y_"-names
colnames(dat2)[grep("x_", colnames(dat))] <- 
                     colnames(dat)[grep("y_", colnames(dat))]
colnames(dat2)
#[1] "TRUE."    "y_value1" "y_value2" "x_value1" "x_value2" "y_id"    
#[7] "x_id"    

rbind(dat,dat2)
#-----------------
  TRUE. x_value1 x_value2 y_value1 y_value2 x_id y_id
1     0        1        4       11       14    1    7
2     1        2        5       12       15    2    8
3     0        3        6       13       16    3    9
4     0       11       14        1        4    7    1
5     1       12       15        2        5    8    2
6     0       13       16        3        6    9    3

Upvotes: 3

FXQuantTrader
FXQuantTrader

Reputation: 6891

If dd is your original dataframe, then you could try:

dd2 <- cbind(dd[1], dd[4:5], dd[2:3],dd[7:6])

names(dd2) <- names(dd)

dd <- rbind(dd, dd2)

Upvotes: 4

Related Questions