BoundsError with Julia DataFrame

General Usage
1

Hello everybody,

I am working on an AI experiment involving generating candidate solutions to a problem. My candidates are in the form of smaller dataframes derived from the larger dataset dataframe. The error I’m facing is this:

ERROR: BoundsError: attempt to access "attempt to access a data frame with 475 rows at index [0, 202, 402, 259]"
Stacktrace:
 [1] getindex(::DataFrames.DataFrame, ::Array{Int64,1}, ::Colon) at /home/jake/.julia/packages/DataFrames/Iyo5L/src/dataframe/dataframe.jl:387
 [2] randomCandidate(::Int64) at /home/jake/Projects/gp-experiment/Experiment.jl:51
 [3] mutate at /home/jake/Projects/gp-experiment/Experiment.jl:28 [inlined]
 [4] main() at /home/jake/Projects/gp-experiment/Experiment.jl:84
 [5] top-level scope at none:0

And this is the abridged version of my code.

function mutate(parent)
    randomCandidate(4)
end

function randomCandidate(n::Integer)
    # Select n random rows from the dataset.
    rows = [randRow() for i = 1:n]
    df[rows, :]
end

function generateInitialPopulation(lambda::Integer, candidateSize::Integer)
    [randomCandidate(candidateSize) for i = 1:lambda]
end

function main()
    # Local variables omitted
    lambda = 4
    mu = 1
    
    while generationNum != GENLIMIT || (best != nothing && fitness(best) != 0)
        # Population fitness evaluation omitted.
        for p in parents
            for i = 1:(lambda/mu)
                push!(pop, mutate(p))
            end
        end

        println("Generation $generationNum, best $best, fitness $bestFitness")
        generationNum += 1
    end

end

Apologies for the awful formatting, randomCandidate works in generateInitialPopulation, but when I try to call it as part of mutate, I get the BoundsError, seemingly nothing has changed, other then where I am calling it from, anyone have any idea what might be happening?

2

Indexing in Julia is 1-based by default. It looks like randRow() (not provided?) is outputting 0’s. The valid integer range for that dataframe is [1, 475].

2 Likes
3

That was it, thank you so much