MLJ- Unpack not defined

General Usage
,
1

Getting an error when attempting to unpack the Boston RDataset

y, X = MLJBase.unpack(boston, ==(Symbol(outcome_name)), colname -> true)
UndefVarError: unpack not defined

Stacktrace:
 [1] getproperty(::Module, ::Symbol) at .\Base.jl:13
 [2] top-level scope at In[6]:1

When I go to look up documentation:

?unpack
search: uncompact

Couldn't find unpack
Perhaps you meant unstack, unlock, unmark, stack or uncompact
No documentation found.

Binding unpack does not exist.

Thoughts? Ive tried updating and re-adding the packages within the MLJ Universe.

MLJ Version=v"0.2.3"
Julia Version 1.3.1

1 Like
2

That’s a pretty old MLJ version (current stable is 0.9), maybe unpack didn’t exist then?

2 Likes
3

I thought it looked low… why would I have such an old version of MLJ if i just installed and updated today?

4

Must be something holding you back - try ]add MLJ@0.9.0 and the error message will tell you

1 Like
5

@Phillip_Sutton apologies, you’re experiencing the pains that I warned (on Slack) users may experience due to a significant refactoring cycle that we just completed…
@nilshg is correct.

The easiest is to rm repos that block the way (i.e. other MLJ* repos) to do so: do ] add MLJBase@0.11 and see what it tells you, remove everything until you can add it. Then add MLJ again. (You should get MLJBase 0.11, MLJ 0.9)

using MLJBase, RDatasets
boston = dataset("MASS", "boston")
y, X = unpack(boston, ==(:Crim), colname -> true)

will return

([0.00632, 0.02731, 0.02729, 0.03237, ...
...
│ Row │ Zn      │ Indus   │ Chas  │ NOx     │ Rm      │ Age     │ Dis     │
│     │ Float64 │ Float64 │ Int64 │ Float64 │ Float64 │ Float64 │ Float64 │
├─────┼─────────┼─────────┼───────┼─────────┼─────────┼─────────┼─────────┤
│ 1   │ 18.0    │ 2.31    │ 0     │ 0.538   │ 6.575   │ 65.2    │ 4.09    │
│ 2   │ 0.0     │ 7.07    │ 0     │ 0.469   │ 6.421   │ 78.9    │ 4.9671  │

(I just tested it)

2 Likes
6

Thanks @tlienart and @nilshg.

The problems in dependencies kept going and going. I’m on a new system and ended up reinstalling julia adn adding MLJ first. Everything seems to be working now.

1 Like