R语言中使用data.table实现分组唯一值求和

• 已定义data.table结构数据集dt，完整结构代码如下：

dt <- structure(list(group = c("A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "A", "A", "A", "A", "A", "A", "A", "A", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B", "B", "B", "B", 
                          "B", "B", "B", "B", "B"), 
               V1 = c(6.38, 6.38, 6.38, 6.38, -1.53, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, -24.93, 
                      -24.93, -24.93, -24.93, -6.8, -6.8, -6.8, -6.8, -6.8, -1.71, 
                      -1.71, -1.71, -1.71, -1.71, -1.06, -1.06, -1.06, -1.06, -1.06, 
                      -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, 
                      -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, 
                      -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, -1.06, 
                      -1.06, -1.06, -1.06, -1.06, -1.06, 8.42, 8.42, 8.42, 4.34, 
                      4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 4.34, 
                      4.34, 4.34, 4.34, 4.34, 4.34, 4.34), 
               V2 = c(0.7, 0.7, 0.7, 0.7, 0.7, 0.7, 0.7, -0.11, -0.11, -11, -11, -11, -11, -11, 
                      -11, -11, -11, -11, -11, 1.6, 1.6, 1.6, 1.6, -0.55, -0.55, -0.55, 
                      -2.15, -2.15, -2.15, -2.15, -2.15, -2.15, -2.15, -2.15, -2.15, 
                      -0.19, -0.19, -0.19, -0.19, -0.19, 2.63, 2.63, 2.63, 2.63, 2.63, 
                      2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 2.63, 
                      -3.86, -3.86, 0.48, 0.48, 0.48, 0.48, 0.48, 0.48, 0.48, -1.38, 
                      -1.38, -1.38, -1.38, -5.15, -11.58, -11.58, -11.58, -11.58, -11.58, 
                      -11.58, -11.58, -11.58, -11.58, -0.46, -7.32, -7.32, -7.32, -7.32, 
                      -7.32, -7.32, -7.32, -7.32, 2.67, 4.88, 4.88, 4.88, 4.88, 4.88, 
                      4.88, 4.88, 4.88, -11.57, -11.57, -11.57, 1.67, 1.55, 1.55, 2.3, 
                      2.3, 2.3, 2.3, 2.3, 2.3, 2.3, 2.3, -1.42, 21.88, 21.88, 21.88, 
                      21.88, 21.88, 21.88, 21.88, 21.88, 21.88, 21.88, 21.88, 21.88, 
                      21.88, 21.88, 21.88, 21.88, 21.88, 21.88, -0.59, -0.59, -0.59, 
                      -0.59, -0.59, -1.87, -1.87, -1.87)), row.names = c(NA, -138L), 
          class = c("data.table", "data.frame"))

• 核心需求：按照group列分组，对V1、V2两列分别计算组内去重后的唯一值总和

现有实现的问题 #

• 原代码直接对全列值求和，重复值会被多次累计，无法得到去重后的结果
• 原代码如下：

library(data.table)
dt[, lapply(.SD, sum, na.rm = T), by=group, .SDcols = c("V1", "V2")]

• 原代码输出结果：

group       V1     V2
1:     A -1571.53 -88.67
2:     B    20.55 245.64

预期输出效果 #

group       V1     V2
1:     A  -269.38 -12.43
2:     B    -4.47  27.17

调整后的正确实现 #

• 核心逻辑：对每一个分组内的目标列，先通过unique()提取不重复值，再执行求和操作
• 调整后代码：

library(data.table)
dt[, lapply(.SD, function(col) sum(unique(col), na.rm = TRUE)), 
   by = group, 
   .SDcols = c("V1", "V2")]

• 代码说明：
  • 保留原有按group分组、通过.SDcols指定计算列的逻辑
  • 替换原有直接调用sum的逻辑，嵌套unique()先做组内去重
  • 保留na.rm = TRUE参数，兼容列内存在空值的场景

内容的提问来源于stack exchange，提问作者Saurabh