GroupBy then combine changes column order

[Tambup]

Hi all,

I just noticed that when groupby and then combine are used, the order of columns changes. I think it is a reasonable behavior, but I think it would be good to point out with a "warning" in the doc. This can lead to unexpected behavior in some cases.

Julia version: 1.9.3
DataFrames version: 1.5.0

Example:

df = DataFrame(A = 1:4, B = ["M", "F", "F", "M"])
df = combine(first, groupby(df, :B))
for (Bval, Aval) in eachrow(df) # I expect here (Aval, Bval) without seeing any warning in the doc
    println("Col B value is:", Bval, "\tCol A value is:", Aval)
end

[bkamins]

Technically the order of columns is not changed (the point is that the columns not specified for grouping are dropped). However, indeed we could add a sentence that if keepkeys=true then in combine grouping columns are put first and are followed by results of the aggregation as I think it is missing.

[Tambup]

Yes, I think it is a great idea.