R脚本分段运行正常但整段运行报`-.POSIXt`错误求助

整段运行R脚本时立即终止并返回如下错误：

Error in `-.POSIXt`(`*tmp*`) : 
  unary '-' is not defined for "POSIXt" objects

但分段运行（例如先执行Q4部分，再执行Q5部分等）时完全正常。相关脚本如下：

# Installing packages I will need

install.packages("dplyr")
install.packages("tidyr")
install.packages("readxl")
install.packages("psych")
install.packages("car")

# loading packages

library(dplyr)
library(tidyr)
library(readxl)
library(psych)
library(car)


# Importing the Data Frame 
aqdata <- read_xlsx("D:\\tVNS - Data Folder\\AnalitiQs\\Company_Y_base_data.xlsx")

# Inspecting the head, colnames and nrows to see that everything was imported successfully
head(aqdata)
colnames(aqdata)
nrow(aqdata)

# Using distinct to compare nrows to ensure that all rows are unique 
distinct(aqdata, Employee_ID)  

--------------
# Q4 Step 1: # Converting the stop Month column to a date vector 
aqdata$Stop_Month <- as.Date(aqdata$Stop_Month)

# Q4 Step 2 - Filtering employees by department & missing values (Q depends on if last working day is date of Stop_Month or (as I interpret) last day of work was 1 day before Stop_Month date)
workers_post28 <- aqdata %>% 
  filter(Department == "Operations" & is.na(Stop_Month))

# Q4 Step 3: Final Answer
num_rows4 <- nrow(workers_post28)
message <- paste("On the 28th of July 2020", num_rows4, "people worked within the Operations Department of company Y")
print(message)
--------------
  
# Q5 Step 1: Converting the Start Month column to a date vector 
  aqdata$Start_Month <- as.Date(aqdata$Start_Month)

# Q5 Step 2: Filtering employees by dep, permanent contract. &  28th of Dec 2019 
workers_on28Dec2019 <- aqdata %>%
  filter(Department == "Operations" & Type_of_Contract == "Permanent") %>%
  
# Q5 Step 3: Filtering resulting employees by condition start date before Dec 28th 2019 and Stop date after 28th of Dec 2019 (based on prev question interpretation)
  filter(Start_Month < as.Date("2019-12-28") & is.na(Stop_Month) | Stop_Month > as.Date("2019-12-28"))

# Q5 Step 4: Final Answer 
num_rows5 <- nrow(workers_on28Dec2019)
messageQ5 <- paste("On the 28th of Dec 2019", num_rows5, "people in the Operations department had a permanent contract")
print(messageQ5)

# Q5 Step 3.1: Because I was unsure if my result was correct, I decided to verify that no employee not satisfying my conditions were left in the dataframe. 

# Filter to only include employees who started working after December 28th, 2019
before_28dec2019 <- workers_on28Dec2019 %>%
  filter(Start_Month > as.Date("2019-12-28"))

# Count the number of rows in the final result
num_rowscheck <- nrow(final_result)

# Print the result
if (num_rowscheck == 0) {
  print("The resulting dataframe does not contain any employees who started working before December 28th, 2019")
}

--------------
  
# Q6 Step 1: Calculate sum of distance and dividing by length of column (Can also use mean() function for more concise code)
AverageDistance <- sum(aqdata$Distance_Work_KM) / length(aqdata$Distance_Work_KM)
  messageQ6 <- paste("The average distance to work for any employee who has ever worked for Company Y is ", AverageDistance, "KM")
  print(messageQ6)
--------------

# Q7 Step 1: Filter workers by gender with conditions satisfying working on 2020-07-28 --> Creating the groups 

Q7df <- aqdata %>%
    
# Step 1.1 Filter rows based on conditions of the question (using same interpretation from Q4 where NA in Stop_Month indicates that they were working on the 28th as this is their last working day)
    filter(Start_Month < as.Date("2020-07-28") & is.na(Stop_Month)) %>%
    
# Step 1.2 Group by gender
    group_by(Gender) %>%
    
# Step 1.3 Select Salary column to incl into subset DF 
    select(Gender,Salary)
    
# Q7 Step 2: Assumption checking first normality, then equality of variances 
describe(Q7df$Salary) # Getting M, SD and Kurtosis values for Normality assumption --> Skew/Kurtosis do not fall within -2 or +2 so its violated 
shapiro.test(Q7df$Salary) # Another way of testing the assumption --> Also indicates violation. 

# Both test indicate that the assumption is violated but our sample is 319 so it should be okay 

# Testing equality of variances 
leveneTest(Q7df$Salary ~ Q7df$Gender) # The test is non-significant meaning that the assumption is not violated 

# Q7 Step 3: Running the t-test 
Outcome <- t.test(Salary ~ Gender, data = Q7df) # Assuming equal variances b.c leven's was non-sig.

# Q7 Step 4: Final conclusions 
print(Outcome$estimate)
print(Outcome$statistic)
print(Outcome$p.value)
messageQ7 <- paste("The T-test indicates that there is significant differences between the means of Female/Male in terms of Salary")
print(messageQ7)

--------------

核心原因 #

脚本中用来分隔各部分的--------------是无效的R语法，整段运行时R会将其解析为连续的一元负号运算符（即反复对前一行代码的执行结果取负）。如果前一行的结果是POSIXt类型的日期对象，就会触发unary '-' is not defined for "POSIXt" objects错误。而分段运行时，你手动跳过了这些分隔线，因此不会触发错误。

另外还有一个隐藏问题：Q5部分的num_rowscheck <- nrow(final_result)中，final_result变量未定义，整段运行到此处也会报错。

解决步骤 #

1. 替换无效分隔线：将所有--------------改为注释格式，比如# ------------------------------，这样R会忽略这些行，不会尝试执行。
2. 修复未定义变量：把Q5中的num_rowscheck <- nrow(final_result)改为num_rowscheck <- nrow(before_28dec2019)，因为before_28dec2019才是前面代码生成的验证用数据框。
3. 统一日期转换逻辑：建议在数据导入后一次性完成日期列的转换，避免分散在各个问题模块重复操作，示例代码如下：

# 统一转换日期列
aqdata <- aqdata %>%
  mutate(
    Stop_Month = as.Date(Stop_Month),
    Start_Month = as.Date(Start_Month)
  )

额外优化建议 #

• 避免重复安装包：可以添加判断逻辑，只在包未安装时才执行安装操作：

# 检查并安装所需包
required_pkgs <- c("dplyr", "tidyr", "readxl", "psych", "car")
for(pkg in required_pkgs) {
  if(!require(pkg, character.only = TRUE)) {
    install.packages(pkg)
    library(pkg, character.only = TRUE)
  }
}

• 模拟全新运行环境：整段运行前清空工作区（RStudio中点击Session -> Clear Workspace），可以提前发现分段运行时因环境变量残留而隐藏的错误。

内容的提问来源于stack exchange，提问作者Boxingday