翻阅古今
使用DataFrame.sort_index:pd.concat([df1, df2], axis=1).sort_index(axis=1)编辑:print (df1) Column col5 col20 1 31 2 4print (df2) Column col10 51 6df = pd.concat([df1, df2], axis=1)c = df.columns.tolist()df = df.reindex(c[:1] + sorted(c[1:]), axis=1)print (df) Column col5 col1 col20 1 5 31 2 6 4EDIT1:与一起使用DataFrame.xs,DataFrame.sort_index添加原始非选定的caolumns值Index.union和最后更改顺序DataFrame.reindex:print (df) Column a col2 col1 col5 col1 col30 1 5 3 5 41 2 6 4 7 7cols = (df.xs('Column', drop_level=False, axis=1, level=0) .sort_index(ascending=False, axis=1).columns)print (cols)MultiIndex([('Column', 'col5'), ('Column', 'col2'), ('Column', 'col1')], )df = df.reindex(cols.union(df.columns, sort=False), axis=1)print (df) Column a col5 col2 col1 col1 col30 3 1 5 5 41 4 2 6 7 7