如何在R中结合case_when条件实现减法与乘法运算

原代码核心问题 #

1. 逻辑错误：计算修正项时误用了当前行的corr1，而非参考行（B/C行）的corr1。比如C行的修正项应该是B行corr1 * B行corr3，但原代码写成了当前行corr1 * B行corr3，导致结果不符合预期。
2. 代码健壮性不足：直接用corr3[colname == "B"]这种列索引方式，若后续数据框行顺序变动，会导致取值错误，可读性也较差。

修正方案 #

方案1：提前提取参考值（推荐） #

先提取B、C行的固定参考值，再执行计算，逻辑清晰且不易出错：

df1 <- data.frame(colname=c("A","B","C","D"),
                  corr1=c(5500, 2400, 3900,5600),
                  corr2=c(2.1944, 2.1936, 2.1934,2.1932),
                  corr3=c(NA, 0.7934, NA ,NA),
                  corr4=c(NA, 0.3104, NA, NA),
                  corr5=c(NA, NA, 0.7933, NA)
)
library(dplyr, warn = FALSE)

# 提取B、C行的参考值
b_corr1 <- df1$corr1[df1$colname == "B"]
b_corr3 <- df1$corr3[df1$colname == "B"]
b_corr4 <- df1$corr4[df1$colname == "B"]
c_corr1 <- df1$corr1[df1$colname == "C"]
c_corr5 <- df1$corr5[df1$colname == "C"]

# 计算最终值
df1 <- df1 %>% 
  mutate(corr_final = case_when(
    colname == "C" ~ corr1*corr2 - (b_corr1 * b_corr3),
    colname == "D" ~ corr1*corr2 - (b_corr1 * b_corr4) - (c_corr1 * c_corr5),
    .default = corr1*corr2
  ))

方案2：dplyr链内完成计算 #

如果希望在一条链式操作中完成，可结合filter和pull提取参考值，最后移除临时列：

df1 <- data.frame(colname=c("A","B","C","D"),
                  corr1=c(5500, 2400, 3900,5600),
                  corr2=c(2.1944, 2.1936, 2.1934,2.1932),
                  corr3=c(NA, 0.7934, NA ,NA),
                  corr4=c(NA, 0.3104, NA, NA),
                  corr5=c(NA, NA, 0.7933, NA)
)
library(dplyr, warn = FALSE)

df1 <- df1 %>% 
  mutate(
    # 临时提取参考值
    b_corr1 = filter(., colname == "B") %>% pull(corr1),
    b_corr3 = filter(., colname == "B") %>% pull(corr3),
    b_corr4 = filter(., colname == "B") %>% pull(corr4),
    c_corr1 = filter(., colname == "C") %>% pull(corr1),
    c_corr5 = filter(., colname == "C") %>% pull(corr5),
    # 计算最终值
    corr_final = case_when(
      colname == "C" ~ corr1*corr2 - (b_corr1 * b_corr3),
      colname == "D" ~ corr1*corr2 - (b_corr1 * b_corr4) - (c_corr1 * c_corr5),
      .default = corr1*corr2
    )
  ) %>% 
  select(-b_corr1, -b_corr3, -b_corr4, -c_corr1, -c_corr5) # 清理临时列

内容的提问来源于stack exchange，提问作者dk1