R data.table 子组计数和组汇总的加权百分比

Question

我有以下data.table

n = 100000

DT = data.table(customer_ID = 1:n,
                married = rbinom(n, 1, 0.4),
                coupon = rbinom(n, 1, 0.15))

我需要创建一个表格，按婚姻状况子组汇总已婚和未婚客户总数、使用优惠券的客户数量，最后一列按婚姻状况计算每个子组使用优惠券的客户百分比。

输出应该是这样的。

   married Customers using Coupons Total Customers percent_usecoupon
1:       0                    9036           59790          15.11290
2:       1                    5943           40210          14.77991

我当前的代码效率非常低，我确信使用 data.table 有更好的语法，但我似乎找不到它。我在下面复制了我当前的代码：

coupon_marital = DT[coupon == TRUE, .N, by = married][order(-N)] #Count of coupon use by marital status
total_marital = DT[, .N, by = married] #Total count by marital status
setnames(total_marital, "N", "Count") #Rename N to Count
coupon_marital = merge(coupon_marital, total_marital) #Merge data.tables

coupon_marital[, percent_usecoupon := N/Count*100, by = married] #Compute percentage coupon use
setnames(coupon_marital, c("N", "Count"), c("Customers using Coupons", "Total Customers")) #Rename N to Count
rm(total_marital)

print(coupon_marital)

我不能使用 dplyr，只需要使用 data.table。我对 data.table 语法相当陌生，非常感谢任何帮助！

Answer 1

创建数据

set.seed(10)
n = 100000
DT = data.table(customer_ID = 1:n,
                married = rbinom(n, 1, 0.4),
                coupon = rbinom(n, 1, 0.15))

汇总数据

DT[, .(N.UseCoupon   = sum(coupon)
      ,N.Total       = .N
      ,Pct.UseCoupon = 100*mean(coupon)), 
   by = married]

#    married N.UseCoupon N.Total Pct.UseCoupon
# 1:       0        8975   60223      14.90294
# 2:       1        5904   39777      14.84275