How to convert all columns of type Int8 to Int64

Specific Domains Data
1

I have a dataframe with quite a few columns of type Int8 that I need to convert to Int64 to satisfy XLSX.jl non-support for integer sizes other than Int64.

Currently I do it column by column:
sf[!,:Enumeration] = convert.(Int64, sf[:, :Enumeration]) #needed by XLSX

There must be a better way, but I am too much of a novice to figure it out.

2

You can find the names of the columns that are Int8 using names(df, Int8).

ulia> using DataFrames

julia> df = DataFrame(a = rand(1:10, 5), b = randn(5), c = Int8.(rand(11:20, 5)))
5×3 DataFrame
 Row │ a      b           c    
     │ Int64  Float64     Int8 
─────┼─────────────────────────
   1 │    10   0.0594745    17
   2 │     5  -2.09879      12
   3 │    10  -0.765042     17
   4 │     6  -0.258906     12
   5 │     7  -0.721766     18

julia> names(df, Int8)
1-element Vector{String}:
 "c"
1 Like
3

First of all, welcome to our community! :confetti_ball:

Complementing the suggestion of @dmbates, use names and then iterate over it:

for int8_col_name in names(sf, Int8)
    sf[!, int8_col_name] = convert.(Int64, sf[:, int8_col_name])
end
4

This is even more concise:

transform!(sf, names(sf, Int8) .=> ByRow(Int), renamecols=false)
3 Likes