TTR runSD returns all NA

Question

I think this came from an R/package update but now when I try to calculate the running standard deviation of a timeseries with an NA in it all I get is NA (current is R version 4.0.3, and TTR_0.24.2)

How do I get:

TTR::runSD(x = c(NA, 1:10), n = 1, cumulative = TRUE)

to exclude / ignore NA instead of returning:

 [1] NA NA NA NA NA NA NA NA NA NA NA

To something more like:

 [1] NA NaN 0.7071068 1.0000000 1.2909944 1.5811388 1.8708287 2.1602469 2.4494897 2.7386128 3.0276504

EDIT ideally return what it did under R version 3.5.2 and TTR_0.23-4 :

> TTR::runSD(x = c(NA, 1:10), n = 1, cumulative = TRUE)
 [1]       NA       NA 1.000000 1.172604 1.414214 1.677051 1.949359 2.226732 2.507133 2.789489 3.073181

It idoesn't have `na.rm`. You may use ` TTR::runSD(x = na.omit(c(NA, 1:10)), n = 1, cumulative = TRUE)` — akrun, Dec 23 '20 at 21:43
hmm this isn't ideal because `na.omit` would change the size of the input/output — Rafael, Dec 23 '20 at 21:44
very good work around, thanks! will wait a while before accepting. holding out to see if there's an in-function solution — Rafael, Dec 23 '20 at 21:51

score 1 · Accepted Answer · answered Dec 23 '20 at 21:47

An option is to create a function and update only those elements having non-NA elements

f1 <- function(vec) {
      i1 <- !is.na(vec)
      vec[i1] <- TTR::runSD(x = vec[i1], n = 1, cumulative = TRUE)
      vec
  }

f1(c(NA, 1:10))
#[1]        NA       NaN 0.7071068 1.0000000 1.2909944 1.5811388 1.8708287 2.1602469 2.4494897 2.7386128 3.0276504

TTR runSD returns all NA

1 Answers1