concat sorts columns without preserving order

[EVaisman]

Although it's my desired behavior, I realize perhaps this is not the desired behavior, but I think it's reasonable to bring up.

If I want to concat two data frames and their column names are the same, the order is preserved:

>>> df1 = pd.DataFrame(columns=["A","1"])
>>> pd.concat([df1,df1])
Empty DataFrame
Columns: [A, 1]
Index: []

However, if they are not, it seems they are sorted:

>>> df2 = pd.DataFrame(columns=["A","2"])
>>> pd.concat([df1,df2])
Empty DataFrame
Columns: [1, 2, A]
Index: []

Would it be reasonable to try to preserve the order? So the output would be:

>>> pd.concat([df1,df2])
Empty DataFrame
Columns: [A, 1, 2]
Index: []

I realize that, in an ordered list, "A" doesn't come before "1" or "2", but in both df1 and df2, "A" does come before "1" and "2".

[jreback]

this is a dupe of #4588

thanks for the report