汇总所有因子水平组合的值，包括缺失值答案

【问题标题】：Aggregate values for all combinations of factor levels including missing ones汇总所有因子水平组合的值，包括缺失值
【发布时间】：2017-08-15 02:52:48
【问题描述】：

我正在尝试根据多列查找数据框的最小值。我可以使用下面的聚合函数成功地做到这一点。但是，结果不包含输入数据框中没有数据的因素组合。

我有什么：

# all possibilities of fruits, cities, and vegetables:
fruits<-c('apple','banana','grape')
cities<-c('new york','chicago','los angeles')
vegetables<-c('cucumber','mushroom')

#my input (ie, a sample from a test:
inputdf<-data.frame(fruit=c('apple','apple','apple','banana','banana','banana','grape','grape','grape'),city=c('new york','new york','new york','new york','chicago','los angeles','chicago','chicago','chicago'),vegetable=c('cucumber','cucumber','mushroom','cucumber','mushroom','mushroom','cucumber','cucumber','cucumber'),value=c(5,3,4,6,5,7,2,7,4))

#my aggregation:
outdf<-aggregate(value ~ fruit + city + vegetable,inputdf,function(x) min(x))

我得到的输出是：

fruit   city        vegetable   value
grape   chicago     cucumber    2
apple   new york    cucumber    3
banana  new york    cucumber    6
banana  chicago     mushroom    5
banana  los angeles mushroom    7
apple   new york    mushroom    4

这是正确的，但是，我还想要与输入 df 中根本不存在的列组合相对应的行：

fruit   city        vegetable   value
apple   new york    cucumber    3
apple   new york    mushroom    4
apple   chicago     cucumber    NA
apple   chicago     mushroom    NA
apple   los angeles cucumber    NA
apple   los angeles mushroom    NA
banana  new york    cucumber    6
banana  new york    mushroom    NA
banana  chicago     cucumber    NA
banana  chicago     mushroom    5
banana  los angeles cucumber    NA
banana  los angeles mushroom    7
grape   new york    cucumber    NA
grape   new york    mushroom    NA
grape   chicago     cucumber    2
grape   chicago     mushroom    NA
grape   los angeles cucumber    NA
grape   los angeles mushroom    NA

我希望能够对要组合的任意数量的列执行此操作。有没有一种简单的方法可以做到这一点？我想要该输出的原因是因为我需要将 NA 转换为特定值并再次在相同子集上平均这些值。谢谢！

【问题讨论】：

标签： r

【解决方案1】：

您可以使用expand.grid生成所有组合，然后使用merge：

outdf<-aggregate(value ~ fruit + city + vegetable,inputdf,function(x) min(x))
DF=expand.grid(fruits, cities, vegetables)
outdf=merge(outdf,DF,by.x=c('fruit','city','vegetable'),by.y=c('Var1','Var2','Var3'),all.y=T) 
> outdf
    fruit        city vegetable value
1   apple     chicago  cucumber    NA
2   apple     chicago  mushroom    NA
3   apple los angeles  cucumber    NA
4   apple los angeles  mushroom    NA
5   apple    new york  cucumber     3
6   apple    new york  mushroom     4
7  banana     chicago  cucumber    NA
8  banana     chicago  mushroom     5
9  banana los angeles  cucumber    NA
10 banana los angeles  mushroom     7
11 banana    new york  cucumber     6
12 banana    new york  mushroom    NA
13  grape     chicago  cucumber     2
14  grape     chicago  mushroom    NA
15  grape los angeles  cucumber    NA
16  grape los angeles  mushroom    NA
17  grape    new york  cucumber    NA
18  grape    new york  mushroom    NA

【讨论】：