Lag function for time series

0
votes

I have a question on the lag function which I am unable to solve in R. I have a variable in a dataframe 'V3' which is a time series of a very large data file. The 'resultV4' is what I want to accomplish (see the code snippet).

  1. I want my resultV4 to be -1 in the first row, based on V3.
  2. In the 2nd row of V3 there is a zero, so I want the value of resultV4 to be -1 (the value of the 1st row)
  3. In the 3rd row of V3 there is a zero, so I want the value of resultV4 to be -1 (the value of the 1st row)

When the value of V3 changes, in this case to 1 in the 6th row, I want resultV4 to be 1. The 7th row is another 0, so I want this to be the value of the 6th row in V3, which is 1. And so on..

V3<-c(-1,0,0,0,0,1,0,0,0,0,-1,0,0,0,0,-1,0,0)
resultV4<-c(-1,-1,-1,-1,-1,1,1,1,1,1,-1,-1,-1)
df<-cbind(V3,resultV4)

Thanks in advance for any suggestions.

Cheers,

PCdL

2
rlag

2 Answers

2
votes

You can use na.locf from package zoo.

library(zoo)

V3 <- c(-1,0,0,0,0,1,0,0,0,0,-1,0,0,0,0,-1,0,0)
V3_adj <- V3

replace 0 with NA 

my_zero <- which(V3 == 0)

V3_adj[my_zero] <- NA

carry forward the last observation

resultV4 <- na.locf(V3_adj)

cbind(V3, V3_adj, resultV4)

Result:

     V3 V3_adj resultV4
 [1,] -1     -1       -1
 [2,]  0     NA       -1
 [3,]  0     NA       -1
 [4,]  0     NA       -1
 [5,]  0     NA       -1
 [6,]  1      1        1
 [7,]  0     NA        1
 [8,]  0     NA        1
 [9,]  0     NA        1
[10,]  0     NA        1
[11,] -1     -1       -1
[12,]  0     NA       -1
[13,]  0     NA       -1
[14,]  0     NA       -1
[15,]  0     NA       -1
[16,] -1     -1       -1
[17,]  0     NA       -1
[18,]  0     NA       -1
1
votes
resultV4 <- V3
for(i in seq_along(V3)) { if(resultV4[i] == 0) resultV4[i] = resultV4[i-1] }
resultV4
[1] -1 -1 -1 -1 -1  1  1  1  1  1 -1 -1 -1 -1 -1 -1 -1 -1