Julia DataFrame ERROR: attempt to access data frame with 3 columns at index [0, 0, 0]-CodePudding

everyone, I am a Julia newbie. I want to write a Function that delete any column with only zeroes.


    function delEmptyCol(df)
    emptyColList = []
    for col in eachcol(df)
        if sum(col) == 0
            append!(empyColList,col)
        end
    end
    newdf = select!(df,Not(emptyColList))
    return newdf
    end

and I made up a trial DataFrame df2 to test my Function.It looks like following.

KFC Mc  Piz
Int64   Int64   Int64
1   0   1   4
2   0   2   5
3   0   3   6

So what I hope to get is as following.

    Mc  Piz
Int64   Int64
1   1   4
2   2   5
3   3   6

However,when i do delEmptyCol(df2),I get an error and I have no idea what is wrong.

BoundsError: attempt to access data frame with 3 columns at index [0, 0, 0]

Stacktrace:
  [1] getindex
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\other\index.jl:199 [inlined]
  [2] getindex
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\other\index.jl:257 [inlined]
  [3] getindex
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\other\index.jl:224 [inlined]
  [4] manipulate(df::DataFrame, c::InvertedIndex{Vector{Any}}; copycols::Bool, keeprows::Bool, renamecols::Bool)
    @ DataFrames C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\abstractdataframe\selection.jl:1680
  [5] #select#492
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\abstractdataframe\selection.jl:1171 [inlined]
  [6] #select!#487
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\abstractdataframe\selection.jl:873 [inlined]
  [7] select!
    @ C:\Users\cxh\.julia\packages\DataFrames\zqFGs\src\abstractdataframe\selection.jl:873 [inlined]
  [8] delEmptyCol(df::DataFrame)
    @ Main .\In[46]:8
  [9] top-level scope
    @ In[51]:1
 [10] eval
    @ .\boot.jl:373 [inlined]
 [11] include_string(mapexpr::typeof(REPL.softscope), mod::Module, code::String, filename::String)
    @ Base .\loading.jl:1196

Please help me !! I'd appreciate!

CodePudding user response：

There might be an easier way than this, but you could use mapcols to get a dataframe with booleans that say false if all elements are 0. Then use this to subset the columns.

julia> df = DataFrame(KFC = [0, 0, 0], Mc = [1,2,3], Piz = [4,5,6])
3×3 DataFrame
 Row │ KFC    Mc     Piz   
     │ Int64  Int64  Int64 
─────┼─────────────────────
   1 │     0      1      4
   2 │     0      2      5
   3 │     0      3      6

julia> df[:, Array(mapcols(col -> any(col .!= 0), df)[1, :])]
3×2 DataFrame
 Row │ Mc     Piz   
     │ Int64  Int64 
─────┼──────────────
   1 │     1      4
   2 │     2      5
   3 │     3      6

I used any and .!= to look for columns that have something other than 0 in them. This gives you a dataframe with a single row, which is then extracted and converted to an array.

CodePudding user response：

An alternative to what @niczky12 proposed is:

julia> select(df, all.(!=(0), eachcol(df)))
3×2 DataFrame
 Row │ Mc     Piz
     │ Int64  Int64
─────┼──────────────
   1 │     1      4
   2 │     2      5
   3 │     3      6

(the condition is a bit different than what @moczly12 proposed as I understand you want to drop a column if all its elements are 0, but this is a detail - I understand you are asking about the general approach)