Yet Another XArray-like Julia Package

Documentation	Build Status

Summary

YAXArrays.jl is a package to handle gridded data that is larger than memory. It enables the DiskArray.jl package to access the data lazily and provides map and mapCube to apply user defined functions on arbitrary subsets of the axes. These computations are also easily parallelized either via Distributed or via Threads.

Switch to DimensionalData

With YAXArrays.jl 0.5 we switched the underlying data type to be a subtype of the DimensionalData.jl types. Therefore the indexing with named dimensions changed to the DimensionalData syntax. See the DimensionalData.jl docs and the `Switch to DimensionalData section in our docs.

Installation

Install the YAXArrays package:

julia>]
pkg> add YAXArrays

You may check the installed version with:

] st YAXArrays

Start using the package:

using YAXArrays

Quick start

Let's assemble a YAXArray with 3 RangeAxis, i.e. time, x,y and a CategoricalAxis with two variables.

axlist = [
    RangeAxis("time", range(1, 20, length=20)),
    RangeAxis("x", range(1, 10, length=10)),
    RangeAxis("y", range(1, 5, length=15)),
    CategoricalAxis("Variable", ["var1", "var2"])]

and the corresponding data.

data = rand(20, 10, 15, 2)

You might also add additional properties via a Dictionary, namely

props = Dict(
    "time" => "days",
    "x" => "lon",
    "y" => "lat",
    "var1" => "one of your variables",
    "var2" => "your second variable",
)

And our first YAXArray is built with:

ds = YAXArray(axlist, data, props)

YAXArray with the following dimensions
time                Axis with 20 Elements from 1.0 to 20.0
x                   Axis with 10 Elements from 1.0 to 10.0
y                   Axis with 15 Elements from 1.0 to 5.0
Variable            Axis with 2 elements: var1 var2 
Total size: 46.88 KB

Getting data back from a YAXArray

For axis can be via .

ds.x.values

1.0:1.0:10.0

and the complete data for one of our variables, i.e. var1 can be accessed via:

ds[variable = "var1"].data

For more please take a look at the documentation.