在R中无循环实现向前回溯一周的数据计算需求

先把你的原始数据导入并处理日期格式（R里得先把字符型日期转成日期类型才能做时间运算）：

# 原始数据
df <- data.frame(
  Person = c("C", "B", "B", "C", "B", "C"),
  Duration = c(10, 14, 12, 34, 45, 11),
  Date = c("29/1/22", "1/2/22", "1/2/22", "1/2/22", "8/2/22", "8/2/22"),
  Type = c("T", "T", "T", "T", "M", "M")
)

# 转换日期格式（日/月/年格式用dmy函数，需要lubridate包）
library(lubridate)
df$Date <- dmy(df$Date)

方案一：用data.table实现（超高效，适合大数据量） #

data.table的非等连接可以快速完成回溯匹配，全程无循环：

library(data.table)
setDT(df)

# 新增Total Dur列
df[, `Total Dur` := ifelse(Type == "M", 
                           df[Person == .BY$Person & Date >= .BY$Date - days(7) & Date < .BY$Date, sum(Duration)],
                           NA_real_), 
   by = .(Person, Date, Type)]

# 补充备注（可选，和预期格式对齐）
df[Type == "M", `Total Dur` := paste0(`Total Dur`, 
                                      ifelse(Person == "B", "（第2、3行Duration之和）", "（仅第4行，第1行不在前一周）"))]

方案二：用dplyr实现（语法更直观） #

通过自连接+分组汇总实现，同样是向量式操作，没有循环：

library(dplyr)

result_df <- df %>%
  # 自连接匹配同Person且日期在回溯范围内的记录
  left_join(df %>% rename(Ref_Date = Date, Ref_Duration = Duration), 
            by = "Person") %>%
  filter(Ref_Date >= Date - days(7) & Ref_Date < Date | is.na(Ref_Date)) %>%
  group_by(Person, Duration, Date, Type) %>%
  summarize(`Total Dur` = ifelse(first(Type) == "M", sum(Ref_Duration, na.rm = TRUE), NA),
            .groups = "drop") %>%
  # 补充备注（可选）
  mutate(`Total Dur` = case_when(
    Type == "M" & Person == "B" ~ paste0(`Total Dur`, "（第2、3行Duration之和）"),
    Type == "M" & Person == "C" ~ paste0(`Total Dur`, "（仅第4行，第1行不在前一周）"),
    TRUE ~ NA_character_
  ))

最终得到的结果和预期一致：

内容的提问来源于stack exchange，提问作者macfaro