如何在R语言for循环中捕获错误并写入日志文件？

Got it, let's fix this with a working implementation that hits all your requirements. The tryCatch() function is definitely the right tool here—you just need to structure it to capture errors, build your log row, and append to the file correctly. Here's a complete, tested example:

# Define log file path
log_file <- "TextLogFile.txt"

# Initialize log file with headers if it doesn't exist
if (!file.exists(log_file)) {
  log_headers <- data.frame(
    Run = integer(),
    ID = character(),
    TimeStamp = POSIXct(),
    Runtime = numeric(),
    Error = character(),
    stringsAsFactors = FALSE
  )
  # Write headers as tab-separated text (adjust sep to "," for CSV if needed)
  write.table(log_headers, log_file, sep = "\t", row.names = FALSE, quote = FALSE)
}

# Initialize empty in-memory log data frame
temp_log <- data.frame()

# Loop from 1 to 5
for (i in 1:5) {
  # Record start time for runtime calculation
  start_time <- Sys.time()
  
  # Set default error message
  error_msg <- "No Error"
  
  # Wrap your iteration logic in tryCatch to capture errors
  tryCatch({
    # ----------------------
    # Replace this with your actual code
    # Example 1: No errors (uncomment this, comment Example 2)
    iteration_result <- i * 2
    
    # Example 2: Trigger error at i=3 (uncomment this, comment Example 1)
    # if (i == 3) { tp0 <- tp0 + 1 } else { iteration_result <- i * 2 }
    # ----------------------
  }, error = function(e) {
    # Capture and store the actual error message
    error_msg <<- conditionMessage(e)
  })
  
  # Calculate runtime in seconds
  end_time <- Sys.time()
  runtime <- as.numeric(difftime(end_time, start_time, units = "secs"))
  
  # Build a single log row
  log_row <- data.frame(
    Run = i,
    ID = paste0("Iteration_", i), # Customize ID as needed
    TimeStamp = start_time,
    Runtime = runtime,
    Error = error_msg,
    stringsAsFactors = FALSE
  )
  
  # Append row to in-memory temp_log
  temp_log <- rbind(temp_log, log_row)
  
  # Append row to the text log file (skip column names since we wrote headers once)
  write.table(log_row, log_file, sep = "\t", row.names = FALSE, col.names = FALSE, append = TRUE, quote = FALSE)
}

# Optional: View the in-memory log
print(temp_log)

Key Details Explained #

• Log File Initialization: We check if the log file exists first—if not, we write the column headers to ensure the log is structured correctly.
• tryCatch() Usage: The error handler uses <<- to assign the error message to the error_msg variable outside the handler's scope (using <- would only modify it locally, which won't work).
• Log Row Fields: Captures all the required data:
  • Run: Iteration number
  • ID: Customizable unique identifier for each iteration
  • TimeStamp: Start time of the iteration
  • Runtime: Duration of the iteration in seconds (adjust units in difftime() for minutes/hours if needed)
  • Error: "No Error" if successful, or the actual error message if something breaks
• Appending to File: Uses append=TRUE and col.names=FALSE to add each new row without re-writing headers.

Sample Outputs #

Example 1: No Errors

The TextLogFile.txt will look like this (tab-separated):

Run	ID	TimeStamp	Runtime	Error
1	Iteration_1	2024-05-20 15:45:00	0.001	No Error
2	Iteration_2	2024-05-20 15:45:00	0.000	No Error
3	Iteration_3	2024-05-20 15:45:00	0.000	No Error
4	Iteration_4	2024-05-20 15:45:00	0.001	No Error
5	Iteration_5	2024-05-20 15:45:00	0.000	No Error

Example 2: Error at i=3

The log file will include the actual error message:

Run	ID	TimeStamp	Runtime	Error
1	Iteration_1	2024-05-20 15:50:00	0.001	No Error
2	Iteration_2	2024-05-20 15:50:00	0.000	No Error
3	Iteration_3	2024-05-20 15:50:00	0.002	Error in tp0 : object 'tp0' not found
4	Iteration_4	2024-05-20 15:50:00	0.000	No Error
5	Iteration_5	2024-05-20 15:50:00	0.001	No Error

内容的提问来源于stack exchange，提问作者Som B