在R中为二元运算符错误提供非数字参数

Question

我正在尝试生成一个变量来标记家庭名册中受访者与其父母之间难以置信的年龄差异。 mapply() 在“二元运算符的非数字参数”上给出错误，而当我仅在两列上应用该函数时，我没有收到此错误。很感谢任何形式的帮助。下面，我试图做一个可重现的例子。

# Variables
respbirth <- c(1974, 1950, 1990, 1980 ) 
B1010_1 <- c(1950, 1960, 1960, 1979 ) 
B1040_1 <- c(3,3,3,3)
B1010_2 <- c(1974, NA, NA, 1975 ) 
B1040_2 <- c(3,1,3,3)

# Data frame
df <- data.frame(respbirth, B1010_1, B1040_1, B1010_2, B1040_2 ) 
df

# Generate empty variable for flaging cases
df$flag_parent <- FALSE

## Generate a function flagging implausible differences using year of birth
attach(df)  # the function doesnt work without this for some reason
imp.parent <- function(data=df,parentAge=B1010_1,relationship=B1040_1) {
  df$flag_parent <- with(df, ((respbirth-parentAge)<18) & (relationship==3))
return(df)
}

# Test
df <- imp.parent(parentAge=B1010_1,relationship=B1040_1)

# Apply this function to all columns
parentAge <- c(paste0("B1010_",1:19, sep=""))
relationship <- c(paste0("B1040_",1:19, sep=""))
mapply(imp.parent, parentAge, relationship )

Answer 1

目前，您的mapply 尝试存在许多问题，包括参数类型、函数调用、返回值等。为避免列出这些问题，请考虑以下重构代码。

# Generate a function flagging implausible differences using year of birth
imp.parent <- function(parentAge, relationship, data=df) {    
  ((df$respbirth - df[[parentAge]]) < 18) & (df[[relationship]] == 3)  
}

# Apply this function to all columns
parentAge <- c(paste0("B1010_", 1:2))
relationship <- c(paste0("B1040_", 1:2))

# Assign columns True/False
df[paste0("false_flag_", 1:2)] <- mapply(imp.parent, parentAge, relationship )

df    
#   respbirth B1010_1 B1040_1 B1010_2 B1040_2 false_flag_1 false_flag_2
# 1      1974    1950       3    1974       3        FALSE         TRUE
# 2      1950    1960       3      NA       1         TRUE        FALSE
# 3      1990    1960       3      NA       3        FALSE           NA
# 4      1980    1979       3    1975       3         TRUE         TRUE

事实上，你甚至需要 mapply（隐藏循环）！ R 可以计算跨等长数据块的逻辑条件，用于列块的矢量化分配：

# Apply this function to all columns
parentAge <- c(paste0("B1010_", 1:2))
relationship <- c(paste0("B1040_", 1:2))

# NOTICE USE OF `[` (NOT `[[`)
df[paste0("false_flag_", 1:2)] <- ((df$respbirth - df[parentAge]) < 18) & (df[relationship] == 3) 

df    
#   respbirth B1010_1 B1040_1 B1010_2 B1040_2 false_flag_1 false_flag_2
# 1      1974    1950       3    1974       3        FALSE         TRUE
# 2      1950    1960       3      NA       1         TRUE        FALSE
# 3      1990    1960       3      NA       3        FALSE           NA
# 4      1980    1979       3    1975       3         TRUE         TRUE

Online Demo