I wanted to get the standard deviation of the 3 previous row of the data, the present row and the 3 rows after.
This is my attempt:
mutate(ming_STDDEV_SAMP = zoo::rollapply(ming_f, list(c(-3:3)), sd, fill = 0)) %>%
Result
| ming_f | ming_STDDEV_SAMP |
|---|---|
| 4.235279667 | 0.222740262 |
| 4.265353 | 0.463348209 |
| 4.350810667 | 0.442607461 |
| 3.864739333 | 0.375839159 |
| 3.935632333 | 0.213821765 |
| 3.802632333 | 0.243294783 |
| 3.718387667 | 0.051625808 |
| 4.288542333 | 0.242010836 |
| 4.134689 | 0.198929941 |
| 3.799883667 | 0.112733475 |
This is what I expected:
| ming_f | ming_STDDEV_SAMP |
|---|---|
| 4.235279667 | 0.225532646 |
| 4.265353 | 0.212776157 |
| 4.350810667 | 0.23658801 |
| 3.864739333 | 0.253399417 |
| 3.935632333 | 0.26144862 |
| 3.802632333 | 0.246259684 |
| 3.718387667 | 0.20514358 |
| 4.288542333 | 0.208578409 |
| 4.134689 | 0.208615874 |
| 3.799883667 | 0.233948429 |
CodePudding user response：
It doesn't match your output exactly, but perhaps this is what you need:
zoo::rollapply(quux$ming_f, 7, FUN=sd, partial=TRUE)
(It also works replacing 7 with list(-3:3).)
This expression isn't really different from your sample code, but the output is correct. Perhaps your original frame has a group_by still applied?
Data
quux <- structure(list(ming_f = c(4.235279667, 4.265353, 4.350810667, 3.864739333, 3.935632333, 3.802632333, 3.718387667, 4.288542333, 4.134689, 3.799883667), ming_STDDEV_SAMP = c(0.225532646, 0.212776157, 0.23658801, 0.253399417, 0.26144862, 0.246259684, 0.20514358, 0.208578409, 0.208615874, 0.233948429)), class = "data.frame", row.names = c(NA, -10L))