带分组的Cumsum表

Question

我怎样才能得到按性别和州分组的Cumsum表？

Gender = sample(c('male', 'female'), 100, replace=TRUE)
State = sample(c('CA', 'WA', 'NV', 'OR', "AZ"), 100, replace=TRUE)
Number = sample(1:8, size=100, replace=TRUE)

df <- data.frame(Gender,State, Number)

Answer 1

如果我们在找累积表，那么

library(data.table)
dcast(setDT(df)[, .N, .(Gender, State, Number)
      ][, perc := round(100*N/sum(N), 2), .(Gender, State)],
     Gender + State ~Number, value.var = 'perc', fill = 0, drop = FALSE)[, 
     (3:10) := lapply(Reduce(`+`, .SD, accumulate = TRUE),
            function(x) paste0(x, "%")), .SDcols = -(1:2)][]

Answer 2

对于更简单的方法，我建议使用dplyr。在加载tidyverse时，Dplyr将与其他一些有用的包一起加载。

library(tidyverse)

Gender = sample(c('male', 'female'), 100, replace=TRUE)
State = sample(c('CA', 'WA', 'NV', 'OR', "AZ"), 100, replace=TRUE)
Number = sample(1:8, size=100, replace=TRUE)

df <- data.frame(Gender,State, Number)

df <- df %>% 
  group_by(Gender, State) %>% 
  mutate(Number_CumSum = cumsum(Number)) %>% 
  ungroup() %>% 
  arrange(State, Gender)

head(df)

# A tibble: 6 x 4
  Gender  State Number Number_CumSum
  <fctr> <fctr>  <int>         <int>
1 female     AZ      8             8
2 female     AZ      3            11
3 female     AZ      4            15
4 female     AZ      5            20
5 female     AZ      2            22
6 female     AZ      7            29