Hi,
I'm processing the data from an oscilloscope. The natural way to
encapsulate them is to have a structure in a form of
type Wfm
xorigin
xinc
len
ydata
properties::Dict{ASCIIString,Any}
end
, where X represents time and Y data are equidistantly spaced in time. To
access a subset one would index into Wfm using the range defined in the X
axis:
function getindex(W::Wfm, I::FloatRange)
function getx(W::Wfm, I::FloatRange)
wfm = wfmopen("file.h5")
x = getx(wfm, -1e-8:1e-8)
y = wfm[-1e-8:1e-8]
plot(x,y)
However, this is not possible. During the FloatRange construction omitting the
stepsize the end value is lost and it is not possible to reconstruct again.
> a = -1e-8:1e-8
> dump(a)
FloatRange{Float64}
start: Float64 -1.0e-8
step: Float64 1.0
len: Float64 1.0
divisor: Float64 1.0
I understand that for array indexing or array construction the current
behavior of FloatRange with default stepsize=1.0 is OK, but it blocks much
more appealing general use-cases not related to array indexing. And the
notation xmin:xmax is appealing and very well readable. Much better that
forcing the user to write -1e-8:1e-16:1e8 or -1e-8:eps(Float64):1e-8.
Do you think an addition of an exact_end field into this type in Base would
be worth the effort and storage space?
Petr
