Shifting multiple columns down by one row in R

Question

I have number of columns in my dataset and I need to shift columns 18:101 down by one row. So far, I found this command in another thread to be helfpul and below I apply it to my data:

data.xts$AR2_Forecast <- transform(data.xts$AR2_Forecast, AR2_Forecast = c(NA, AR2_Forecast[-nrow(data.xts$AR2_Forecast)]))

However, since I have 83 additional columns, changing the names would be quite time-consuming. Is there anything I could use that would save me the time?

I was thinking of something along the lines of:

data.xts[18:101]<-

but I am not sure how the transform function should look. The class of my data is "xts", "zoo".

Thanks for any feedback!

The same approach also works with the `xts` class. See my edited answer. — user12728748, Apr 01 '20 at 22:59

score 0 · Answer 1 · answered Apr 01 '20 at 16:13

0

I think you can use dplyr::select() along with dplyr::lag() to do this more efficiently.

answered Apr 01 '20 at 16:13

Nikhil Gupta

1,436
12
15

user12728748 · Answer 2 · 2020-04-01T20:16:17.883

0

You could use the same approach as for one column, e.g. by doing:

df[, 18:101] <- rbind(rep(NA, 101-17), df[seq_len(nrow(df)-1), 18:101] )

Edit: For xts, try:

df[, 18:101] <- rbind(rep(NA, 101-17), coredata(df[1:(nrow(df)-1), 18:101] ))

edited Apr 01 '20 at 20:16

answered Apr 01 '20 at 17:38

user12728748

8,106
2
9
14

This method doesn't work, as far as I understand it is because the class of my data is xts, not data.frame – Adam Uhliarik Apr 01 '20 at 19:35

Shifting multiple columns down by one row in R

2 Answers2