CODEWITHSUNDEEP
rdataframe
user3356873
user3356873

Reputation: 63

Use value from the previous row for manipulation of multiple column at once in R

Hi everyone I need a help.
I have the data-set similar to this which contain multiple rows and columns. 

df<- data.frame(A=seq(0.01,0.05,0.01),
            B=c(0.01, -0.24, 0, -0.21, 0), 
            C=seq(0.03,0.07,0.01),
            D=c(0.4,0.5,0,0,2))

I used shift command and created another row E. 

df[ , E := shift(A)+A]

Now I want to apply similar function to whole data frame df and create row F, G, H similar to E using similar method at once.

Thank you.

Upvotes: 0

Views: 139

Answers (2)

akrun
akrun

Reputation: 887128

If we are using data.table, the shift can take multiple columns at once and output the lag of those. If we are not selecting any particular sets of columns, specifying shift(.SD) (.SD represents the Subset of Data.table) gives the lag of all the columns in the dataset. By assigning (:=) it to different column names (LETTERS[5:8]), we get a new set of lag columns in the original dataset.

library(data.table)
setDT(df)[, LETTERS[5:8] := shift(.SD)+.SD]
df
#     A     B    C   D    E     F    G   H
#1: 0.01  0.01 0.03 0.4   NA    NA   NA  NA
#2: 0.02 -0.24 0.04 0.5 0.03 -0.23 0.07 0.9
#3: 0.03  0.00 0.05 0.0 0.05 -0.24 0.09 0.5
#4: 0.04 -0.21 0.06 0.0 0.07 -0.21 0.11 0.0
#5: 0.05  0.00 0.07 2.0 0.09 -0.21 0.13 2.0

Or we can loop through lapply

setDT(df)[, LETTERS[5:8] := lapply(.SD, function(x) x+shift(x))]

Upvotes: 3

Kunal Puri
Kunal Puri

Reputation: 3427

Here is an alternative for this:

new_cols <- c('E','F','G','H')

old_cols <- colnames(df)

for (i in seq_along(new_cols)){

    eval(parse(text = paste0("df[,",new_cols[i],":= shift(",old_cols[i],")+",old_cols[i],"]")))
}

Upvotes: 1

Related Questions