2025年Julia - HDF5、JLD2库、group、dataset

科技前沿 • 2025-03-15 16:17 • 阅读 29

大家好，我是讯享网，很高兴认识大家。

JLD2库中，group和dataset是什么样的安排，取决于读、写的细节。

一、准备h5材料

using H5JL; using DelimitedFiles; using JLD2; csv_file = "C:\\Users\\songroom\\Desktop\\000001.XSHE.csv"; jld_file = "C:\\Users\\songroom\\Desktop\\000001.h5"; # readdlm header=true,=>tuple @time csv_data,~ = DelimitedFiles.readdlm(csv_file,',',header =true); # @time :6.5 seconds new_csv = H5JL.drop_array_null(csv_data); #不用管，去空格的 # 从csv=>h5_array h5_data = H5JL.get_h5_array_from_csv_by_fullcode(new_csv,"000001.XSHG"); println("h5_data::$(size(h5_data))") #@time h5_data = turn_larger(h5_data,20) group_dataset_name1 = "1000.XSHG/000001.XSHG"; group_dataset_name2 = "1000.XSHG/000002.XSHG"; # write @time H5JL.write_h5_data(jld_file,h5_data,group_dataset_name1) # 4.13s @time H5JL.write_h5_data(jld_file,h5_data,group_dataset_name2) # 4.13s

讯享网

好了，我们就生成了一个.h5文件。

二、测试JLD2的h5结构

讯享网jld_file = "C:\\Users\\songroom\\Desktop\\000001.h5"; if isfile(jld_file) f = JLD2.jldopen(jld_file, "r+")# "r+"; else f = JLD2.jldopen(jld_file, "w")# "r"; "w"=>覆盖了 end

1、判断文件中的基本属性

julia> isempty(f) false julia> f.writable #文件是否可“写” true julia> f.written true julia> f.compress false julia> f.mmaparrays false julia> f.n_times_opened 1

2、group名字

讯享网julia> keys(f.root_group) 1-element Array{String,1}: "1000.XSHG"

3、创建新group

julia> JLD2.Group(f, "1001.XSHG") JLD2.Group (no datasets)

4、dataset名字

讯享网

讯享网julia> datasets = keys(f["1000.XSHG"]) 2-element Array{String,1}: "000001.XSHG" "000002.XSHG"

5、dataset取值

julia> f["1000.XSHG"]["000002.XSHG"] ×3 Array{Float64,2}: 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ⋮ 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

讯享网julia> JLD2.Group(f, "1000.XSHG/000003")# no datasets是指下面没有dataset JLD2.Group (no datasets) julia> keys(f.root_group) 1-element Array{String,1}: "1000.XSHG" julia> keys(f["1000.XSHG"]) 3-element Array{String,1}: "000001.XSHG" "000002.XSHG" "000003"

julia> f["1000.XSHG/000004"] =4 4 julia> keys(f["1000.XSHG"]) # 查看dataset已经变化 4-element Array{String,1}: "000001.XSHG" "000002.XSHG" "000004" "000003"

注意：但不能在原有的基础进行修改！

讯享网julia> f["1000.XSHG/000004"] =5 ERROR: ArgumentError: a group or dataset named 000004 is already present within this group

（3）另一个创建：要求是新的group!

 julia> JLD2.Group(f, "0004.XSHG")["stock1"] = 600 600

在这里插入图片描述这个就说明，目前阶段，dataset不能重新赋值。如果需要重新赋值，则应先把其中的数据备份下来，删除原文件，再新建！

8、keys

讯享网dataset1 = "1000.XSHG/000001.XSHG" dataset2 = "1000.XSHG/000002.XSHG" dataset3 = "1001.XSHG" haskey(f.root_group, dataset1) haskey(f.root_group, dataset2) haskey(f.root_group, dataset3)

2025年Julia - HDF5、JLD2库、group、dataset

相关推荐