CODEWITHSUNDEEP
r
Wicelo
Wicelo

Reputation: 2426

Substract one element from another in next row but different column?

Basically I'm looking for a diff() function that works on dataframes and that can substract one element in row n from another element in row n+1 but that is also in another column.

So, with the dataframe : 

> df = data.frame(X=c("x1","x2","x3"), Y=c("y1","y2","y3"))
   X  Y
1 x1 y1
2 x2 y2
3 x3 y3

I would like to do df$Z=df$Y-df$Z but with a "lag" so I get :

> df$Z
[1] NULL y2-x1 y3-x2

Without using a loop of course, I've never seen such a slow thing such as this in R...

Upvotes: 0

Views: 88

Answers (2)

akrun
akrun

Reputation: 887241

Try (You showed some factor columns instead of numeric values). If the columns are numeric,

 df$Z <- c(NA,df$Y[-1]- df$X[-nrow(df)])

Or using with

 df$Z <-   with(df, c(NA,Y[-1]- X[-length(X)]))

Explanation

For the data provided

 df$Y[-1]  #removes the first observation
 #[1] y2 y3
 #Levels: y1 y2 y3
 df$X[-nrow(df)] #removes the last observation
 #[1] x1 x2
 #Levels: x1 x2 x3

so that, when we take the difference, it will be y2-x1, y3-x2

Upvotes: 0

MrFlick
MrFlick

Reputation: 206263

Now sure why you had to make things difficult in your example by using character values but

c(NA, with(df, paste(tail(Y,-1), head(X,-1), sep="-")))

show the subtractions you want to perform. if the data were actually numeric

df$Z <- c(NA, with(df, tail(Y,-1) - head(X,-1)))

should do the trick

Upvotes: 1

Related Questions