用dplyr計算多變量分組時變量的比例

Question

我有一個包含列組的tibble，賬戶和持續時間，每行代表1個事件。我想創建一個很好的彙總表，其中包括組，賬戶，總計持續時間，計算價格以及最終總持續時間的組別比例。

重複的樣品：

library(tidyverse) 
library(lubridate) 
tidy_data <- structure(list(group = c("Group 1", "Group 2", "Group 3", "Group 1", "Group 2", "Group 3", "Group 4", "Group 4", "Group 2"), account = c("Account 1", "Account 2","Account 3", "Account 1", "Account 2", "Account 3", "Account 4", "Account 4", "Account 2"), duration = structure(c(146.15, 181.416666666667, 96.9, 52.2833333333333, 99.4333333333333, 334.116666666667, 16.6333333333333, 11.5666666666667, 79.5666666666667), units = "mins", class = "difftime")), class = c("tbl_df", "tbl", "data.frame"), row.names = c(NA, -9L), .Names = c("group","account", "duration")) 
hourPrice = 25 

總結1 - 正確地計算的比例，但不包括帳號

tidy_data %>% 
    group_by(group) %>% 
    summarise(total = sum(duration) %>% time_length(unit = "hour") %>% round(digits = 2), 
         price = (total*hourPrice) %>% round(digits = 0)) %>% 
    mutate(prop = (price/sum(price) * 100) %>% round(digits = 0)) 

# A tibble: 4 × 4 
    group total price prop 
    <chr> <dbl> <dbl> <dbl> 
1 Group 1 3.31 83 20 
2 Group 2 6.01 150 35 
3 Group 3 7.18 180 42 
4 Group 4 0.47 12  3 

摘要2 - 包括帳號，但無法計算比例正確

tidy_data %>% 
    group_by(group, account) %>% 
    summarise(total = sum(duration) %>% time_length(unit = "hour") %>% round(digits = 2), 
         price = (total*hourPrice) %>% round(digits = 0)) %>% 
    mutate(prop = (price/sum(price) * 100) %>% round(digits = 0)) 

#Source: local data frame [4 x 5] 
#Groups: group [4] 

    group account total price prop 
    <chr>  <chr> <dbl> <dbl> <dbl> 
1 Group 1 Account 1 3.31 83 100 
2 Group 2 Account 2 6.01 150 100 
3 Group 3 Account 3 7.18 180 100 
4 Group 4 Account 4 0.47 12 100 

我意識到問題是，由於這兩個在第二種情況下，總結只能在一個組內進行。我考慮完成摘要1，然後將帳號重新加入表格，但在我看來，必須有更好的解決方案。

編輯：輸出我想：

group account total price prop 
    <chr>  <chr> <dbl> <dbl> <dbl> 
1 Group 1 Account 1 3.31 83 20 
2 Group 2 Account 2 6.01 150 35 
3 Group 3 Account 3 7.18 180 42 
4 Group 4 Account 4 0.47 12  3 

Answer 1

相反的summarise，我們使用mutate數據集中創建新列，然後slice每個「組」的第一行，計算「道具'並刪除'持續時間'列

tidy_data %>% 
     group_by(group) %>% 
     mutate(total = sum(duration) %>% 
       time_length(unit = "hour") %>% 
       round(digits = 2), 
       price = (total*hourPrice) %>% 
       round(digits = 0)) %>% 
     slice(1L) %>% 
     ungroup() %>% 
     mutate(prop = (price/sum(price) * 100) %>% 
      round(digits = 0)) %>% 
     select(-duration)  
# A tibble: 4 × 5 
#  group account total price prop 
#  <chr>  <chr> <dbl> <dbl> <dbl> 
# 1 Group 1 Account 1 3.31 83 20 
# 2 Group 2 Account 2 6.01 150 35 
# 3 Group 3 Account 3 7.18 180 42 
# 4 Group 4 Account 4 0.47 12  3