"Un-melt" 数据框并保留其余列? Python Pandas

"Un-melt" Dataframe and keep rest of columns? Python Pandas

我有一个table这种格式,我想用"opposite"融化。还有另一个问题可以解决这个问题,但它不适用于我想保留的许多其他专栏。

原文:

COUNTRY   STATE     CATEGORY   RESTAURANT         STARS     REVIEWS
US        Texas     NaN        Texas Chicken      4.1       1,157    
US        Texas     Spicy      Texas Chicken      4.1       1,157
US        Ohio      NaN        Mamas Shop         3.6       700
US        Ohio      NaN        Pizza Hut          4.5       855
US        Ohio      Pizza      Pizza Hut          4.5       855 

期望的输出:

COUNTRY   STATE     RESTAURANT        STARS    REVIEWS  SPICY   PIZZA 
US        Texas     Texas Chicken     4.1      1,157    1       0 
US        Ohio      Mamas Shop        3.6      700      0       0
US        Ohio      Pizza Hut         4.5      855      0       1 

基本上我想 "group by" 许多列,同时根据类别列中的类别创建额外的列。对于所有这些附加列,没有任何特定类别的餐厅的值为 0。我也不想要任何额外的列层,因为我打算将其全部写入 [​​=27=]。

非常感谢对此的任何帮助,并在此先感谢您!

set_index, crosstab and reindex 的组合可以 'unmelt' 数据帧,并处理数据帧中存在的空值:

#set aside required multiindex of country, state, restaurant, stars, and reviews
ind = df.set_index(['COUNTRY','STATE','RESTAURANT','STARS','REVIEWS']).index

#get frequency count for Pizza and Spicy
res = pd.crosstab([df.COUNTRY,df.STATE,df.RESTAURANT,df.STARS,df.REVIEWS],df.CATEGORY)

#reindex frequency dataframe with ind
res = res.reindex(ind,fill_value=0).drop_duplicates()
res


                CATEGORY                    Pizza   Spicy
COUNTRY STATE   RESTAURANT     STARS  REVIEWS       
 US     Texas   Texas Chicken   4.1    1,157    0   1
        Ohio    Mamas Shop      3.6    700      0   0
                Pizza Hut       4.5    855      1   0

我想这应该可行:

pd.crosstab([df.COUNTRY,df.STATE,df.RESTAURANT,df.STARS,df.REVIEWS], df['CATEGORY'].fillna('_')).drop(columns='_')