CODEWITHSUNDEEP
savejuliaload
MogaGennis
MogaGennis

Reputation: 107

Save data in Julia

Is there a standard way to easily just save any kind of data in Julia. This time I have a tuple with two arrays in it that I want to use again later. Something like this.

a = [1,2,34]
b = [1,2,3,4,5]

save((a,b), "ab")
a, b = load("ab")

Upvotes: 6

Views: 8377

Answers (2)

Cameron Bieganek
Cameron Bieganek

Reputation: 7684

You can use JLD2.jl to save and load arbitrary Julia objects. This is more likely to work for long term storage across different Julia versions than the Serialization standard library.

using JLD2

x = (1, "a")
save_object("mytuple.jld2", x)

julia> load_object("mytuple.jld2")
(1, "a")

For additional functionality, check out the docs.

Upvotes: 11

Przemyslaw Szufel
Przemyslaw Szufel

Reputation: 42234

There are many scenarios depending on how long you want to store the data

  1. Serialization (this is perfect for short time periods, for longer periods where packages update their internal data structures you might have problem reading the file)
julia> using Serialization

julia> serialize("file.dat",a)
24

julia> deserialize("file.dat")
3-element Vector{Int64}:
  1
  2
 34
  1. Delimited files (note that additional processing of output might be needed)
julia> using DelimitedFiles

julia> writedlm("file.csv", a)

julia> readdlm("file.csv", '\t' ,Int)
3×1 Matrix{Int64}:
  1
  2
 34
  1. JSON (great for long term)
julia> using JSON3

julia> JSON3.write("file.json",a)
"file.json"

julia> open("file.json") do f; JSON3.read(f,Vector{Int}); end
3-element Vector{Int64}:
  1
  2
 34

Other worth mentioning libraries (depending on data format) include CSV.jl for saving data frames and BSON.jl for binary JSON files.

Upvotes: 6

Related Questions