Shift a column in R-CodePudding

Suppose that I have a dataframe like the following:

library(tidyverse)

df <- tibble(x = c(1,2,3), y = c(4,5,6))

# A tibble: 3 x 2
      x     y
  <dbl> <dbl>
1     1     4
2     2     5
3     3     6

And I would like to shift a column, adding a column like:

# A tibble: 3 x 3
      x     y shifted_x
  <dbl> <dbl>     <dbl>
1     1     4        NA
2     2     5         1
3     3     6         2

Basically, I want to work with timeseries, so I would like to get previous values and use them as feature. In python I know I can do:

for i in range(1,11):
    df[f'feature_{i}']=df['sepal_length'].shift(i)

Where df is a pandas Dataframe. Any equivalent code for doing this in R?

CodePudding user response：

An equivalent in R tidyverse is dplyr::lag. Create the column in mutate and update the object by assigning (<-) back to the same object 'df'

library(dplyr)
df <- df %>%
    mutate(shifted_x = lag(x))

or if we need to use the shift, there is shift in data.table

library(data.table)
setDT(df)[, shifted_x := shift(x)]

Also, if we need to create more than one column, shift can take a vector of values in n

setDT(df)[, paste0('shifted', 1:3) := shift(x, n = 1:3)]

CodePudding user response：

We can also use the flag function from the collapse package.

library(collapse)

df %>%
  ftransform(shifted_x = flag(x))
# # A tibble: 3 x 3
#       x     y shifted_x
# * <dbl> <dbl>     <dbl>
# 1     1     4        NA
# 2     2     5         1
# 3     3     6         2