如何通过 R 函数运行数据帧的子集？

Question

我正在尝试设置一个函数，该函数对来自我的主数据框的特定子集的选择变量（data.var1，data.var2）进行回归分析，但该函数在整个数据框上运行，而不是不仅仅是我想要的子集，无论我是在函数内部还是外部定义子集。

#Function subsetting data by temp and running regression
varReg21C <- function(data.var1,data.var2) {
  data21C <- subset(allPursuit,allPursuit$temp == 21)
  fitData <- lm(data.var1 ~ data.var2, data21C)
  regData <- summary(fitData)
  anovaData <- anova(fitData)
  reg21C <- list(fitData=fitData,regData=regData,anovaData=anovaData)
}

#OR

#Function running regression on data already in subset
data21C <- subset(allPursuit,allPursuit$temp == 21)
data21C
data25C <- subset(allPursuit,allPursuit$temp == 25)
data25C
data29C <- subset(allPursuit,allPursuit$temp == 29)
data29C

varReg21C <- function(data.var1,data.var2) {
  fitData <- lm(data.var1 ~ data.var2, data21C)
  regData <- summary(fitData)
  anovaData <- anova(fitData)
  reg21C <- list(fitData=fitData,regData=regData,anovaData=anovaData)
}

Answer 1

您的环境中很可能有 data.var1 和 data.var2 浮动。您需要定义所有输入并防止直接从环境中读取，所以如果我们这样写：

varReg <- function(data.var1,data.var2,DataFrame,TempChoice) {
  data <- subset(DataFrame,temp == TempChoice)
  Form <- paste(data.var1,"~",data.var2)
  fitData <- lm(Form, data)
  regData <- summary(fitData)
  anovaData <- anova(fitData)
  return( list(fitData=fitData,regData=regData,anovaData=anovaData))
}

allPursuit = data.frame(x=runif(100),y=runif(100),z=runif(100),
temp=sample(c(21,25,29),100,replace=TRUE))

varReg("x","y",allPursuit,21)[[1]]

Call:
lm(formula = Form, data = data)

Coefficients:
(Intercept)            y  
     0.3126       0.1494 

varReg("x","y",allPursuit,25)[[1]]

Call:
lm(formula = Form, data = data)

Coefficients:
(Intercept)            y  
    0.55069      0.01734