Adding more data to a JLD2 file

General Usage
1

I’m slow with these things, but I cannot get a JLD2 file to open and save new information.

Using the instructions here

jldopen("example.jld2", "w") do file
    file["bigdata"]
end

Is file supposed to be the name of the file, or is it just supposed to be file?
Is "bigdata" meant to be the name of the vector/array/thing you are saving?

I’ve tried it two different ways, but no file is created and nothing is saved

jldopen("test", "a+") do file
    one_foodweb_result2
end

f = jldopen("test.jld2", "a+")
write(f, "test", one_foodweb_result2)
close(f)

What is incorrect with my code? I am using Julia 1.5 in Atom.

2

The following code does what you probably want:

using JLD2
mediumdata=randn(5)
@save "example.jld2" mediumdata # create a file with some data

jldopen("example.jld2", "a+") do file
    file["bigdata"] = randn(6) 
    file["smalldata"] = randn(4) # put your data here instead of random
end
# append data

mediumdata=nothing # reset

@load "example.jld2" bigdata mediumdata smalldata

@show bigdata
@show mediumdata
@show smalldata  # here we are
3

Thank you for your reply. Is it possible this may not work when it is within a function?

4 
using JLD2

function append_jld2(fname, varname, data)
   jldopen(fname, "a+") do file
     file[varname] = data
   end
end

mediumdata=randn(5)
@save "example.jld2" mediumdata # create file

append_jld2("example.jld2", "bigdata", randn(6))


mediumdata=nothing # reset

@load "example.jld2" bigdata mediumdata

@show bigdata
@show mediumdata # here we are
1 Like
5

Thank you. When I rerun code to add more data, I get

ArgumentError: a group or dataset named one_foodweb_result2 is already present within this group

The file is already made, but I don’t understand why it wouldn’t be able to add more data. This is happening within a function, so the data I want to add is not in the global scope. Perhaps the issue is with adding something to the same object that is saved in the file?

6

Yes, to my knowledge you can’t “add” anything to an already saved dataset.