R custom running function

Question

How can I achieve same results without loop? It's equivalent of scan function in reactiveX but with custom logic (http://reactivex.io/rxjs/img/scan.png)

for (i in 2:nrow(x)) {
  x$running_number[i] <-
    ifelse(abs(x$running_number[(i - 1)] - x$numbers[i])  > max_diff,
           x$numbers[i],
           x$running_number[(i - 1)])
}

EDIT:

Not sure I asked correctly after your reply. Take a look at this dataframe

x <- data.frame(
  c(1000, 1010, 1020, 1030, 1100, 1120, 1140, 1150, 1200),
  c(1000, 1000, 1000, 1000, 1100, 1100, 1100, 1100, 1200)
)

colnames(x) <- c('input_v', 'output_v')

Given x$input_v need to get x$output_v

I can express condition like :

x$output_v <-
  ifelse(
    abs(x$input_v - somehow_get_me_last_assigned_x_output_v) > 90,
    x$input_v,
    somehow_get_me_last_assigned_x_output_v
  )

akrun · Accepted Answer · 2019-08-08 13:42:47Z

1

We can use diff to find the difference between the adjacent elements, create the logical condition with 'max-diff', use that in ifelse to return either the 'numbers' or the lag of 'running_number'

with(x, ifelse(c(FALSE, abs(diff(running_number)) > max_diff), 
      numbers, c(NA, running_number[-nrow(x)])))

Or with dplyr

library(dplyr)
x %>%
  mutate(running_number = case_when(c(FALSE, abs(diff(running_number)) > 
             max_diff)0 ~ numbers, TRUE ~ lag(running-number)))

answered Aug 8, 2019 at 13:42

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

akrun Over a year ago

@AndrewVorobyov. i think the logic with max_diff which is 70 may not work in the example you updated. YOu can try

with(x, ave(input_v, cumsum(c(TRUE, diff(input_v) > 20)), FUN = function(x) x[1])) [1] 1000 1000 1000 1000 1100 1100 1100 1100 1200

Collectives™ on Stack Overflow

R custom running function

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related