Getting slope in R using Loops and conditionals

Question

I am working on a data frame with x and y columns with values as rows.. I want to calculate the slope of x and y for every 2 rows and then using the calculated slope, record whether slope's "stability" is "high" or "low". You'll understand better after seeing the code. What is wrong with this code? When I input stability, R returns NULL.

slope <- (acc$y[i+1] - acc$y[i]) / (acc$x[i+1] - acc$x[i])
stability <- c()
for (i in 1:nrow(acc)) {
  if (slope[i] > 0 & slope[i] < 0.8) {
    stability[i] <- "low"
  } else if (slope[i] >= 0.8 & slope[i] <= 1) {
    stability[i] <- "high"
  } else if (slope[i] > 1 & slope[i] < 1.2) {
    stability[i] <- "high"
  } else if (slope[i] >= 1.2) {
    stability[i] <- "low"
  } else if (slope[i] >= -1 & slope[i] <= -0.8) {
    stability[i] <- "high"
  } else if (slope[i] >= -0.8 & slope[i] <= 0) {
    stability[i] <- "low"
  } else if (slope[i] < -1 & slope[i] > -1.2) {
    stability[i] <- "high"
  } else 
    stability[i] <- "low"
}

Please make a reproducible example: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — Jeppe Olsen, Jul 07 '17 at 22:54
When `i` equals `nrow(acc)` what is `acc$y[i+1]`... You may have to change the `for` loop to iterate only in the range`1:(nrow(acc)-1)`. — kangaroo_cliff, Jul 07 '17 at 23:58
As @M.Hassan has done, it is better to do calculate the `slope` inside the loop. — kangaroo_cliff, Jul 08 '17 at 00:17
Calculate all slopes at once using diff(acc$y)/diff(acc$x). Then combine your conditions with | and use the ifelse function. You don't need a loop at all. — Roland, Jul 08 '17 at 00:28

M.Hassan · Answer 1 · 2017-07-08T01:50:19.350

0

Modify your code as following:

    #slope <- (acc$y[i+1] - acc$y[i]) / (acc$x[i+1] - acc$x[i])
     stability <- as.vector(nrow(acc))       
    n <- nrow(acc)-1
    for (i in 1:n) {
      slope <- (acc$y[i+1] - acc$y[i]) / (acc$x[i+1] - acc$x[i])         

      if (slope > 0 && slope < 0.8) {
        stability[i] <- "low"
      } else if (slope >= 0.8 && slope <= 1) {
        stability[i] <- "high"
      } else if (slope > 1 && slope < 1.2) {
        stability[i] <- "high"
      } else if (slope >= 1.2) {
        stability[i] <- "low"
      } else if (slope >= -1 && slope <= -0.8) {
        stability[i] <- "high"
      } else if (slope >= -0.8 && slope <= 0) {
        stability[i] <- "low"
      } else if (slope < -1 && slope > -1.2) {
        stability[i] <- "high"
      } else
        stability[i] <- "low"
    }

Try it Online

edited Jul 08 '17 at 01:50

answered Jul 08 '17 at 00:05

M.Hassan

10,282
5
65
84

You need to have `()` over `nrow(acc) - 1`. As it is now, `i` starts with `0`, not 1. – kangaroo_cliff Jul 08 '17 at 00:15
I set the counter to n <- nrow(acc)-1, thanks for hint – M.Hassan Jul 08 '17 at 01:52
@ Disco Superfly , n starts with 1, bez for (i in 1:n), try the demo – M.Hassan Jul 08 '17 at 12:38
That is because the code is edited now. See your previous comment where you thanked me for the hint... – kangaroo_cliff Jul 09 '17 at 00:39
@ Disco Superfly, I mean by my comment that loop start with 1 bez it's declared in the body of "for" statement, see my old revisions: https://stackoverflow.com/revisions/44981210/1 , https://stackoverflow.com/revisions/44981210/2. I modified the code (nrow(acc)-1) to be enclosed between braces to avoid error, and -1 affect the end of the loop, and I'm still thank you:) – M.Hassan Jul 09 '17 at 11:49

Getting slope in R using Loops and conditionals

1 Answers1