简化R语言批量处理育龄女性数据集的冗余代码需求

问题背景 #

现有一份包含近50万育龄女性的数据集：

• ageownchild_pernum1至ageownchild_pernum30：记录女性所育子女的年龄，未生育的对应字段为NA（例如有2个子女的女性仅前两个字段有值，其余为NA）
• F_curve_notch_X与f_curve_tilde_X：对应15至49.25的年龄段（X为具体年龄值，如15、15.25等）

需要对所有ageownchild_pernum(1:30)及F_curve_notch(15至49.25)执行重复逻辑，手动编写将产生大量冗余代码，需寻求简化方案。

样本数据 #

library("tidyverse")
DataSet1 <- tibble(
  id = c(1,2,3,4,5,6,7,8,9,10),
  ageownchild_pernum1 = c(18,24,13,16,9,NA,17,13,32,7 ),
  ageownchild_pernum2=  c(16,NA,9 ,10,7,NA,13,11,20,5 ),
  AGE=  c(38,52 ,41 ,43 ,38 ,36 ,40 ,36 ,56,31),
  F_curve_notch_15= c(25.14,30.28,43.33,43.33,25.14,25.14,25.14,25.14,67   ,33.77),
  F_curve_notch_15.25= c(34.01,40.33,51.74,51.74,34.01,34.01,34.01,34.01,73.85,41.91),
  f_curve_tilde_15= c(25.14,30.28,43.33,43.33,25.14,25.14,25.14,25.14,67   ,33.77),
  f_curve_tilde_15.25= c(25.14,30.28,43.33,43.33,25.14,25.14,25.14,25.14,67   ,33.77)
)

原手动执行逻辑示例（单字段） #

以下是针对ageownchild_pernum1和F_curve_notch_15的处理逻辑，需批量扩展到所有目标字段：

DataSet1$low_notch      <- ifelse((DataSet1$ageownchild_pernum1>=0), DataSet1$AGE - DataSet1$ageownchild_pernum1 - 0.75, 0)
DataSet1$high_notch     <- ifelse((DataSet1$ageownchild_pernum1>=0), DataSet1$AGE - DataSet1$ageownchild_pernum1 + 0.75, 0)
DataSet1$low_low_notch  <- ifelse((DataSet1$ageownchild_pernum1>=0), DataSet1$AGE - DataSet1$ageownchild_pernum1 - 1.25, 0)
DataSet1$high_high_notch<- ifelse((DataSet1$ageownchild_pernum1>=0), DataSet1$AGE - DataSet1$ageownchild_pernum1 + 1.25, 0)
DataSet1$low_low_notch  <- ifelse((DataSet1$low_low_notch>=20) & (DataSet1$low_low_notch<35), DataSet1$low_low_notch+0.25, DataSet1$low_low_notch)
DataSet1$high_high_notch<- ifelse((DataSet1$high_high_notch>=20) & (DataSet1$high_high_notch<35), DataSet1$high_high_notch+0.25, DataSet1$high_high_notch)

notch <- function(a, b,c,d){
  ifelse((a<= 15)&(b>=15)&(c!= 0), 0.01*d, c)
}
DataSet1$f_curve_notched_15 <- mapply('notch', DataSet1$low_low_notch, DataSet1$high_high_notch, 
                                       DataSet1$f_curve_notched_15, DataSet1$f_curve_tilde_15, DataSet1$f_curve_notched_15)

简化解决方案 #

利用tidyverse的tidyr和dplyr工具，将宽格式数据转换为长格式批量处理，避免重复代码：

步骤1：处理子女年龄字段，批量计算notch变量 #

# 转换子女年龄字段为长格式
child_age_long <- DataSet1 %>%
  select(id, AGE, starts_with("ageownchild_pernum")) %>%
  pivot_longer(
    cols = starts_with("ageownchild_pernum"),
    names_to = "child_num",
    values_to = "child_age",
    values_drop_na = TRUE  # 过滤未生育的NA记录
  ) %>%
  # 计算各类notch变量
  mutate(
    base_val = AGE - child_age,
    low_notch = ifelse(child_age >= 0, base_val - 0.75, 0),
    high_notch = ifelse(child_age >= 0, base_val + 0.75, 0),
    low_low_notch = ifelse(child_age >= 0, base_val - 1.25, 0),
    high_high_notch = ifelse(child_age >= 0, base_val + 1.25, 0),
    # 调整low_low_notch和high_high_notch的值
    low_low_notch = ifelse(low_low_notch >=20 & low_low_notch <35, low_low_notch + 0.25, low_low_notch),
    high_high_notch = ifelse(high_high_notch >=20 & high_high_notch <35, high_high_notch + 0.25, high_high_notch)
  )

步骤2：处理F_curve字段，批量应用notch函数 #

# 转换F_curve字段为长格式
f_curve_long <- DataSet1 %>%
  select(id, starts_with("F_curve_notch"), starts_with("f_curve_tilde")) %>%
  pivot_longer(
    cols = starts_with(c("F_curve_notch", "f_curve_tilde")),
    names_to = c(".value", "age_group"),
    names_pattern = "(.*)_(.*)"  # 拆分字段名，提取前缀和年龄段
  ) %>%
  # 关联之前计算的notch变量
  left_join(child_age_long %>% select(id, low_low_notch, high_high_notch), by = "id") %>%
  # 应用notch逻辑
  mutate(
    F_curve_notch = ifelse((low_low_notch <=15) & (high_high_notch >=15) & (F_curve_notch !=0), 0.01*f_curve_tilde, F_curve_notch)
  ) %>%
  # 转回宽格式（按需保留原结构）
  pivot_wider(
    names_from = age_group,
    values_from = c(F_curve_notch, f_curve_tilde),
    names_glue = "{.value}_{age_group}"
  )

步骤3：合并处理后的数据集 #

# 合并子女notch变量和处理后的F_curve字段
final_data <- DataSet1 %>%
  select(-starts_with(c("F_curve_notch", "f_curve_tilde"))) %>%
  left_join(f_curve_long, by = "id") %>%
  left_join(child_age_long, by = c("id", "AGE"))

说明 #

• 长格式转换后，所有重复逻辑只需编写一次，自动应用到所有目标字段
• 若需要保留原宽格式的子女notch变量，可在child_age_long基础上用pivot_wider转回宽格式
• 针对50万行的大样本，tidyverse的向量化操作效率优于循环或mapply，适合批量处理

内容的提问来源于stack exchange，提问作者Azam Mirzaei