Pretty print tables with more than 2 dimensions

[diegozea]

Right now, 3 dimensional contingency tables are difficult to read:

julia> tbl = dataset("MASS","survey");

julia> freqtable(tbl, :Sex, :Exer, :Smoke)
2x3x4 NamedArray{Int64,3}:
 Sex: ASCIIString["Female" "Male"] Exer: ASCIIString["Freq" "None" "Some"] Smoke: ASCIIString["Heavy" "Never" "Occas" "Regul"][3 0 2
 4 1 1]

[39 10 50
 47 8 34]

[5 1 3
 7 2 1]

[2 0 3
 7 1 4]

julia> R"table($tbl$Sex, $tbl$Exer, $tbl$Smoke)"
WARNING: RCall.jl Warning in bin + pd * (as.integer(cat) - 1L) :
  longitud de objeto mayor no es múltiplo de la longitud de uno menor
Warning in bin + pd * (as.integer(cat) - 1L) :
  longitud de objeto mayor no es múltiplo de la longitud de uno menor
RCall.RObject{RCall.IntSxp}
, ,  = Heavy


         Freq None Some
  Female    3    0    6
  Male      2    0    0

, ,  = Never


         Freq None Some
  Female   45   10   39
  Male     49    9   38

, ,  = Occas


         Freq None Some
  Female    2    1    6
  Male      6    1    3

, ,  = Regul


         Freq None Some
  Female    4    0    3
  Male      4    3    3

[nalimilan]

Yeah, I've noticed that too. But that's nothing to do with FreqTables, it's a NamedArray issue. Please file an issue there, if possible with a reproducer which doesn't use FreqTables for simplicity.