cleaning up if/else r function for previous row value reference-CodePudding

I have a function that is currently working, but I think there may be a better way for it to work without having to manipulate the data so much beforehand. Basically, I am returning a simple TRUE or FALSE if a value in my column is greater than both the two values before it, and after it.

y1 #a single vector column of values
for (i in 3:length(y1)){ #for every number starting at 3 (because for 2 and 1 you can't go back two)
    if(y1[i] > y1[i-1] && y1[i] > y1[i-2] && y1[i] > y1[i 1] && y1[i] > y1[i 2]){ #if the number is greater than 2 before and 2 after...
        y2[i] <- 'TRUE' #if it is true, write true. Here y2[i] you're saving the results in the blank vector
    } else {
        y2[i] <- 'FALSE' } #opposite here
  print(y2[i])

This works okay, but as you see I have to start at 3 in my for loop because otherwise I get an error, given that the first and second values, as well as the last two, can't compute the [i-1],[i-2] or [i 1] and [i 2]. If I do for i:length(y1) it will not work and I also have to add two zeros onto the dataset in order to not get an error/be able to "compute" the last TRUE/FALSE value.

Is there any way to clean up the actual function so that I don't have to manipulate the data beforehand? Essentially have the function give me a null just for the first two and last two values in my data?

CodePudding user response：

Another approach is using lag and lead from dplyr:

library(dplyr)
v2 <- (lag(v1, 2) < v1 & lag(v1, 1) < v1 & lead(v1, 1) < v1 & lead(v1, 2) < v1)

Output & data:

v1 <- c(1,2,3,2,1,1,1,1,1,2,1,1)
v2
[1] FALSE FALSE  TRUE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE FALSE FALSE

CodePudding user response：

my first quick tip would be to look into the lead and lag functions of dplyr. See for example this tutorial or in the dplyr documentation or in Hadley Wickhams R for Data Science. Hope this helps!