测试变量向量并在表上求和，在R中创建新列

Posted 2021-05-06

我有这样一张桌子：

df <- read.table(text = 
                "  Day      city    gender     week
                 'day1'    'city1'   'M'       'one'
                 'day2'    'city2'   'M'       'two'
                 'day1'    'city3'   'F'       'two'
                 'day2'    'city4'   'F'       'two'", 
                 header = TRUE, stringsAsFactors = FALSE) 

我正在计算这样的汇总表：

daily_table <- setDT(df)[, .(Daily_Freq = .N,
                             men = sum(gender == 'M'),
                             women = sum(gender == 'F'),
                             city1 = sum(city == 'city1'),
                             city2 = sum(city == 'city2'),
                             city3 = sum(city == 'city3'),
                             city4 = sum(city == 'city4'),
                             city5 = sum(city == 'city5'))
                         , by = .(week,Day)]

制作这张桌子：

   week  Day Daily_Freq men women city1 city2 city3 city4 city5
    one day1          1   1     0     1     0     0     0     0
    two day2          2   1     1     0     1     0     1     0
    two day1          1   0     1     0     0     1     0     0

但是因为我有几个城市，所以我想使用带有名字的矢量：

cities <- c("city1","city2","city3","city4","city5")

请注意，我的向量中有5个城市，即使其中一个城市的出现次数为零，我希望它出现在我的最终表格中。我该怎么做？

答案

为了确保R显示city5，即使没有对该值的观察，请将其添加为因子级别：

setDT(df)

df[, city :=  factor(city,
                     levels = c("city1","city2","city3","city4","city5"))]

为了避免为每个级别的city写出测试，你可以迭代city的级别，如下所示：

daily_table <- df[, c(.(Daily_Freq = .N,
                        men = sum(gender == 'M'),
                        women = sum(gender == 'F')),
                      lapply(setNames(levels(city), levels(city)),
                             function(x) sum(city == x))),
                  by = .(week,Day)]
daily_table
##    week  Day Daily_Freq men women city1 city2 city3 city4 city5
## 1:  one day1          1   1     0     1     0     0     0     0
## 2:  two day2          2   1     1     0     1     0     1     0
## 3:  two day1          1   0     1     0     0     1     0     0