在 R 中操作数据框（由玩家聚合）答案

【问题标题】：Manipulating data frame in R (aggregating by the player)在 R 中操作数据框（由玩家聚合）
【发布时间】：2017-07-20 17:58:18
【问题描述】：

我有 csv 文件，格式如下：

Player    Sports      Win     Loss
Brian     Football     5       3
Brian     Basketball   4       1
Brian     Bowling      7       0
Chris     Football     3       3
Chris     Basketball   3       4
. . . . 
. . . .

我想将格式更改为以下：

Name&Sports   Win         Loss    Total
Brian         16           4       20
Football      5            3       8
Basketball    4            1       5
Bowling       7            0       7
Chris         6            7       13
Football      3            3       6
Basketball    3            4       7   
. . . .
. . . .

基本上，在新格式中，我们首先写下此人的姓名以及该人参加的所有运动的赢、输、比赛总数。在接下来的行中，我们写下了这个人参加的每项运动，以及在该特定运动中所参加的赢、输、比赛的总数。一旦我们为那个人写了所有东西，我们就会转移到下一个人并做同样的事情。

在 R 中有一种简单的方法吗？

【问题讨论】：

标签： r csv dataframe datatable

【解决方案1】：

df <- read.table(text = "Player    Sports      Win     Loss
Brian     Football     5       3
                 Brian     Basketball   4       1
                 Brian     Bowling      7       0
                 Chris     Football     3       3
                 Chris     Basketball   3       4",header=T)

tmp <- aggregate(df$Win,by=list(df$Player),sum)
tmp <- cbind(tmp, aggregate(df$Loss,by=list(df$Player),sum)[2])
names(tmp) <- colnames(df)[2:4]

df <- rbind(df[,2:ncol(df)], tmp)          
df$Total <- df$Loss + df$Win
df

      Sports Win Loss Total
1   Football   5    3     8
2 Basketball   4    1     5
3    Bowling   7    0     7
4   Football   3    3     6
5 Basketball   3    4     7
6      Brian  16    4    20
7      Chris   6    7    13

或者，如果匹配示例中的行顺序很重要：

df <- rbind(tmp[1,], df[1:3,2:ncol(df)], 
            tmp[2,], df[4:nrow(df),2:ncol(df)]) # could easily be made more programmatic          
df$Total <- df$Loss + df$Win
df

       Sports Win Loss Total
1       Brian  16    4    20
2    Football   5    3     8
3  Basketball   4    1     5
4     Bowling   7    0     7
21      Chris   6    7    13
41   Football   3    3     6
5  Basketball   3    4     7

【讨论】：

【解决方案2】：

来自tidyverse 的解决方案。 dt_final 是最终输出。

# Create example data frame
dt <- read.table(text = "Player    Sports      Win     Loss
Brian     Football     5       3
Brian     Basketball   4       1
Brian     Bowling      7       0
Chris     Football     3       3
Chris     Basketball   3       4",
                 header = TRUE, stringsAsFactors = FALSE)

# Load package
library(tidyverse)

# Split data frame by players
dt_list <- split(dt, f = dt$Player)

# Define a funciton to process data
sum_fun <- function(dt){
  playername <- unique(dt$Player)

  dt1 <- dt %>% 
    mutate(Total = Win + Loss) %>%
    select(-Player) 
  dt2 <- data_frame(Sports = playername,
                    Win = sum(dt1$Win),
                    Loss = sum(dt1$Loss),
                    Total = sum(dt1$Total))
  dt3 <- bind_rows(dt2, dt1)

  return(dt3)
}

# Apply the function
dt_final <- dt_list %>%
  map_df(sum_fun) %>%
  bind_rows() %>%
  rename(`Name&Sports` = Sports)

【讨论】：