JuliaML/MLDatasets.jl

Problem loading OGB dataset ogbg-ppa

chrisn-pik opened this issue · 2 comments

I can not load the OGBDataset: "ogbg-ppa"

using MLDatasets, DataFrames
data = OGBDataset("ogbg-ppa")

but get the following error:

ERROR: MethodError: no method matching parse_pystring(::Missing)
Closest candidates are:
  parse_pystring(::AbstractString) at ~/.julia/packages/MLDatasets/bg0uc/src/utils.jl:2
Stacktrace:
 [1] (::MLDatasets.var"#411#412")(r::DataFrameRow{DataFrame, DataFrames.Index})
   @ MLDatasets ./none:0
 [2] iterate
   @ ./generator.jl:47 [inlined]
 [3] Dict{String, Any}(kv::Base.Generator{DataFrames.DataFrameRows{DataFrame}, MLDatasets.var"#411#412"})
   @ Base ./dict.jl:105
 [4] read_ogb_metadata(fullname::String, dir::Nothing)
   @ MLDatasets ~/.julia/packages/MLDatasets/bg0uc/src/datasets/graphs/ogbdataset.jl:160
 [5] OGBDataset(fullname::String; dir::Nothing)
   @ MLDatasets ~/.julia/packages/MLDatasets/bg0uc/src/datasets/graphs/ogbdataset.jl:132
 [6] OGBDataset(fullname::String)
   @ MLDatasets ~/.julia/packages/MLDatasets/bg0uc/src/datasets/graphs/ogbdataset.jl:131
 [7] top-level scope
   @ REPL[8]:1

I am using Julia 1.8.2, DataFrames v1.3.6 and MLDatasets v0.7.9

@chrisn-pik could you check if the issue is fixed on master?

@CarloLucibello Thanks! I am not getting the error message anymore, however I run out of memory. I tried it on two computers, and it did not work. Are 16GB of RAM not enough to use this dataset? My computer crashes once the download is finished, when trying to extract the data.