Lagged differences

Question

Sample data:

Date <- as.Date(c('1-01-2008','2-01-2008', '3-01-2008','4-01-2008', '5-01-2008', '1-01-2008','2-01-2008', '3-01-2008','4-01-2008', '5-01-2008'), format = "%m-%d-%Y") 
Country <- c('US', 'US','US','US', 'US', 'JP', 'JP', 'JP', 'JP', 'JP') 
Category <- c('Apple', 'Apple', 'Apple', 'Apple', 'Apple', 'Apple', 'Apple', 'Apple', 'Apple', 'Apple') 
Value <- c(runif(10, -0.5, 10))
df <- data.frame(Date, Country, Category, Value)

I am using the following piece to calculate the lagged growth rate of Value within Country and within Category:

df <- ddply(df, .(Country, Category), transform,
                 Growth6m=c(NA, NA, NA, exp(diff(log(Value), lag = 3))-1))

Now I am trying to get lagged difference rather than growth rate. This worked fine for the first lag (i.e. subtracting value from the previous row) like this:

 df <- ddply(df, .(Country, Category), transform,
          Growth1m=c(NA, diff(Value))) 

but when I introduce higher order lags (ex. subtracting the first row from the third row), I get the error such as: "arguments imply differing number of rows: 157, 158". I tried playing around with the NA's but to no avail.

Edit: sample data

Answer 1

That's easy with dplyr

library(dplyr)
df %>%
  group_by(Country, Category) %>%
  mutate(
    deltaLag1 = Value - lag(Value, 1),
    deltaLag2 = Value - lag(Value, 2)
  )