在不同数据帧中提取和减去行 R 的更有效方法

Question

我正在处理这个篮球比赛数据，其中包含大约 50,000 行的数据框游戏。我正在尝试比较每支球队（A 和 B）在每场比赛中的统计数据。

我有另一个名为 teamStats 的数据框，其中有大约 3000 行，每行包含每个赛季的球队。

到目前为止，我已经组装了如下代码：

    for (i in 1:nrow(games)) {
  if (length(which(((teamStats$Year == games$Season[i])==1) & (teamStats$teamID == games$teamA[i]))) == 1) {
    selectTeamA <- teamStats[which(((teamStats$Year == games$Season[i])==1) & (teamStats$teamID == games$teamA[i])),4:45]
  } else {
    selectTeamA <- as.numeric(rep(NA, ncol(differences)))
  }
  if (length(which(((teamStats$Year == games$Season[i])==1) & (teamStats$teamID == games$teamB[i]))) == 1) {
    selectTeamB <- teamStats[which(((teamStats$Year == games$Season[i])==1) & (teamStats$teamID == games$teamB[i])),4:45]
  } else {
    selectTeamB <- as.numeric(rep(NA, ncol(differences)))
  }

  differences[i,] <- selectTeamA - selectTeamB
}

基本上，此代码在对正确的赛季进行子集化后，为每个团队 A 和 B 搜索正确的 teamID。由于每个赛季的每支球队都没有出现在球队统计中，所以我现在用 NA 填补了缺失的行。 “差异”数据框是一个空数据框，它将填充我在 for 循环中的团队 A 和 B 的统计数据差异。

为了让您了解数据：

游戏 - 前 6 行

           Season teamA teamB winner scoreA scoreB
108123   2010  1143  1293      A     75     70
108124   2010  1198  1314      B     72     88
108125   2010  1108  1326      B     60    100
108126   2010  1107  1393      B     43     75
108127   2010  1143  1178      A     95     61

teamStats - 前 6 行和仅前 6 列用于空间 - 完整数据框中具有不同统计信息的大量列。代码为teamID找到正确的行，然后减去G W L等stat列

              School Year teamID  G  W  L
1  abilene christian 2018   1101 32 16 16
2          air force 2018   1102 31 12 19
3              akron 2018   1103 32 14 18
4        alabama a&m 2018   1105 31  3 28
5 alabama-birmingham 2018   1412 33 20 13

结束这篇很长的帖子，我的问题。我的 for 循环代码有效并填充了差异数据框。问题是运行此代码需要 20-30 分钟。我对处理这么多数据不是很有经验。有什么我不知道的技术吗？如何以更有效的方式重写此代码？

Answer 1

一种方法是合并games 和teamStats，作为跨行迭代的替代方法。

复制您的设置的一些代码，以创建一个最小的工作示例：

library(dplyr)
library(purrr)

set.seed(123)
n_games <- 50000
n_teams <- 400
n_years <- 10

games <- data.frame(Season = rep(2005:(2005 + n_years - 1),
                                 each = n_games / n_years)) %>%
  mutate(teamA = sample(1000:(1000 + n_teams - 1), n_games, r = TRUE),
         teamB = map_int(teamA, ~sample(setdiff(1000:(1000 + n_teams - 1), .), 1)),
         scoreA = as.integer(rnorm(n_games, 80, 20)),
         scoreB = as.integer(rnorm(n_games, 80, 20)),
         scoreB = ifelse(scoreA == scoreB, scoreA + sample(c(-1, 1), n_games, r = TRUE), scoreB),
         winner = ifelse(scoreA > scoreB, "A", "B"))

gen_random_string <- function(...) {
  paste(sample(c(letters, " "), rpois(1, 10), r = TRUE), collapse = "")
}

schools_ids <- data.frame(teamID = 1000:(1000 + n_teams - 1)) %>%
  mutate(School = map_chr(teamID, gen_random_string))

teamStats <- data.frame(Year = rep(2005:(2005 + n_years - 1), each = 300)) %>%
  mutate(teamID = as.vector(replicate(n_years, sample(schools_ids$teamID, 300))),
         G = 32, W = rpois(length(teamID), 16), L = G - W) %>%
  left_join(schools_ids)

我们有 50k 行的 games 和 3k 行的 teamStats。现在，我们通过Year 和teamID 将teamStats 折叠成一个小标题：

teamStats <- teamStats %>%
  group_by(Year, teamID) %>%
  nest()

# # A tibble: 3,000 x 3
#     Year teamID data            
#    <int>  <int> <list>          
#  1  2005   1321 <tibble [1 x 4]>
#  2  2005   1192 <tibble [1 x 4]>
#  3  2005   1074 <tibble [1 x 4]>
# <snip>

制作一个小的方便函数来计算差异：

calculate_diff <- function(x, y) {
  if (is.null(x) | is.null(y)) {
    data.frame(G = NA, W = NA, L = NA)
  } else {
    x[, 1:3] - y[, 1:3]
  }
}

现在，我们 (1) 将 games 与 teamStats 连接（或合并），(2) 使用连接的数据集计算差异，以及 (3) unnest（或取消折叠）数据框。

start <- Sys.time()
differences <- games %>%
  left_join(teamStats, c("Season" = "Year", "teamA" = "teamID")) %>%
  rename(teamA_stats = data) %>%
  left_join(teamStats, c("Season" = "Year", "teamB" = "teamID")) %>%
  rename(teamB_stats = data) %>%
  mutate(diff = map2(teamA_stats, teamB_stats, calculate_diff)) %>%
  select(Season, teamA, teamB, diff) %>%
  unnest()
difftime(Sys.time(), start)
# Time difference of 11.27832 secs

结果

head(differences)
#   Season teamA teamB  G  W  L
# 1   2005  1115  1085 NA NA NA
# 2   2005  1315  1177 NA NA NA
# 3   2005  1163  1051  0 -9  9
# 4   2005  1353  1190  0 -4  4
# 5   2005  1376  1286 NA NA NA
# 6   2005  1018  1362  0 -1  1

Answer 2

这是一种使用 tidyverse 包的方法，我希望它应该比 OP 中的循环解决方案快得多。速度（我期望）来自于更多地依赖数据库连接操作（例如 base merge 或 dplyr 的 left_join）来连接两个表。

library(tidyverse)

# First, use the first few columns from the `games` table, and convert to long format with
#   a row for each team, and a label column `team_cat` telling us if it's a teamA or teamB.
stat_differences <- games %>%
  select(row, Season, teamA, teamB)  %>% 
  gather(team_cat, teamID, teamA:teamB) %>%  

# Join to the teamStats table to bring in the team's total stats for that year
  left_join(teamStats %>% select(-row),    # We don't care about this "row"
            by = c("teamID", "Season" = "Year")) %>%

# Now I want to reverse the stats' sign if it's a teamB. To make this simpler, I gather
#   all the stats into long format so that we can do the reversal on all of them, and 
#   then spread back out.
  gather(stat, value, G:L) %>%
  mutate(value = if_else(team_cat == "teamB", value * -1, value * 1)) %>%
  spread(stat, value) %>%

# Get the difference in stats for each row in the original games table.
  group_by(row) %>%
  summarise_at(vars(G:W), sum)

# Finally, add the output to the original table
output <- games %>% 
  left_join(stat_differences)

为了测试这一点，我更改了给定的示例数据，以便两个表相互关联：

games <- read.table(header = T, stringsAsFactors = F,
  text = "row           Season teamA teamB winner scoreA scoreB
108123   2010  1143  1293      A     75     70
108124   2010  1198  1314      B     72     88
108125   2010  1108  1326      B     60    100")

teamStats <- read.table(header = T, stringsAsFactors = F,
  text = "row   School Year teamID  G  W  L
1  abilene_christian 2010   1143 32 16 16
2          air_force 2010   1293 31 12 19
3              akron 2010   1314 32 14 18
4        alabama_a&m 2010   1198 31  3 28
5 alabama-birmingham 2010   1108 33 20 13
6       made_up_team 2018   1326 160 150 10    # To confirm getting right season
7       made_up_team 2010   1326 60 50 10"
)

然后我得到以下输出，这似乎是有道理的。 （我刚刚意识到我应用的收集/变异/传播改变了列的顺序；如果我有时间，我可能会尝试使用 mutate_if 来保持顺序。）

> output
     row Season teamA teamB winner scoreA scoreB   G  L   W
1 108123   2010  1143  1293      A     75     70   1 -3   4
2 108124   2010  1198  1314      B     72     88  -1 10 -11
3 108125   2010  1108  1326      B     60    100 -27  3 -30