PiecewiseIncreasingRanges

PiecewiseIncreasingRanges.jl provides a PiecewiseIncreasingRange type that corresponds to a set of concatenated, monotonically increasing ranges. It implements indexing as well as searchsortedfirst, searchsortedlast, and a findnearest function. This type is useful for holding the potentially discontinuous timestamps corresponding to large quantities of sampled data, where creating and indexing a vector would be undesirably slow.

You can construct a PiecewiseIncreasingRange as below:

julia> PiecewiseIncreasingRange(StepRange{Int,Int}[0:5:15, 18:2:20])
6-element PiecewiseIncreasingRanges.PiecewiseIncreasingRange{Int64,StepRange{Int64,Int64},Void}:
  0
  5
 10
 15
 18
 20

PiecewiseIncreasingRanges also accept an optional divisor, which is divided out of all elements:

julia> PiecewiseIncreasingRange(UnitRange{Int}[0:3, 15:16], 4)
6-element PiecewiseIncreasingRanges.PiecewiseIncreasingRange{Float64,UnitRange{Int64},Int64}:
 0.0 
 0.25
 0.5 
 0.75
 3.75
 4.0

findnearest(rg, x, within_half_sample) finds the index of the element of rg closest to x. If x is equidistant between two samples, it chooses the earlier one. If within_half_sample is true, findnearest throws a NoNearestSampleError if there is no sample within half of a step of x.