redshift中表的当前结构：
year month scenario sourcecountry destinationcountry ags_marketplaceid corestore category pg valuetype main_country_name region_rollup segment_rollup retail_x ba_x fn_x total_x retail_y ba_y fn_y total_y
2017 FEB A 1 INDIA 234 Store None A Forecast India Asia-Pacific H 935.826 0 0 935.826 21.423 0 0 21.423
2017 MAR A 3 OMAN 3 Core None Books Forecast ROW Africa / Middle East & Other C 164.111 0 0 164.111 4 0 0 4
我尝试用相同的方式编写python代码。
-df.setindex（'year'）＃无法设置索引，因为我要在10个以上的字段上设置索引，并将渠道分为零售，ba，fn，总计四部分。不止于此。
df.columns=pd.MultiIndex.from_tuples([tuple(c.split('_',1)) for c in df.columns])
df.stack(0).reset_index(1)
year month scenario sourcecountry destinationcountry ags_marketplaceid corestore category pg valuetype main_country_name region_rollup segment_rollup Channel x y
2017 FEB A 1 INDIA 234 Store None A Forecast India Asia-Pacific H Retail 935.826 21.423
2017 MAR A 3 OMAN 3 Core None Books Forecast ROW Asia-Pacific H BA 164.111 0
2017 MAR A 3 OMAN 3 Core None Books Forecast ROW Africa / Middle East & Other H FN 164.111 0
2017 MAR A 3 OMAN 3 Core None Books Forecast ROW Africa / Middle East & Other H Total 1264.048 21.423