Useful for filling in short runs of missing values in a time series.
Arguments
- x
A double or integer vector of with missing values to fill in using linear interpolation.
- span
A whole number of the maximum span of missing values to interpolate. If a gap exceeds the span none of the values are interpolate.
- tails
A flag specifying whether to fill in missing values at the start and end by setting them to be the same value as the closest adjacent non-missing value.
Details
interpolate_numeric_vector() is essentially a wrapper on stats::approx().
Examples
interpolate_numeric_vector(c(1, NA, 4))
#> [1] 1.0 2.5 4.0
interpolate_numeric_vector(c(1L, NA, 4L))
#> [1] 1 2 4
interpolate_numeric_vector(c(1, NA, NA, NA, NA, 3))
#> [1] 1 NA NA NA NA 3
interpolate_numeric_vector(c(1, NA, NA, NA, NA, 3), span = 4)
#> [1] 1.0 1.4 1.8 2.2 2.6 3.0
interpolate_numeric_vector(c(NA, NA, 10, 1, NA))
#> [1] NA NA 10 1 NA
interpolate_numeric_vector(c(NA, NA, 10, 1, NA), tails = TRUE)
#> [1] 10 10 10 1 1