使用tbl_summary()函数时出现报错，请求技术协助

Hey Lays, sorry you hit this frustrating encoding snag with tbl_summary()! That "invalid multibyte string" error almost always stems from unreadable non-ASCII characters or encoding mismatches in your dataset. Let's walk through some practical fixes to get you back on track:

• Hunt down and clean invalid characters
  First, pinpoint which columns have problematic values using this snippet (replace your_data with your actual data frame):

  library(purrr)
  map(your_data, function(col) which(!validUTF8(col)))
  

  This will show you row indices in character columns with invalid UTF-8 characters. To clean them up, you can either convert non-ASCII characters to their closest ASCII equivalents:

  library(dplyr)
  library(stringi)
  
  your_data <- your_data %>%
    mutate(across(where(is.character), ~stri_trans_general(., "Latin-ASCII")))
  

  Or strip out invalid characters entirely using iconv:

  your_data <- your_data %>%
    mutate(across(where(is.character), ~iconv(., from = "UTF-8", to = "UTF-8", sub = "")))
  

• Validate your data import encoding
  If you loaded your data from a CSV or text file, incorrect encoding during import is a common culprit. When reading the file, explicitly specify UTF-8 encoding:

  your_data <- read.csv("your_file.csv", fileEncoding = "UTF-8")
  

  For Windows users, some files include a byte order mark—try fileEncoding = "UTF-8-BOM" instead if the above doesn't resolve the issue.

• Test with a small dataset subset
  Run tbl_summary() on a tiny slice of your data (like the first 10 rows) to see if the error persists:

  tbl_summary(your_data[1:10, ])
  

  If this works, the problem is isolated to specific rows or columns—you can use this to narrow down exactly where the bad characters live.

• Update the gtsummary package
  Outdated package versions might have unpatched encoding bugs. Update to the latest version with:

  update.packages("gtsummary")
  

If none of these steps fix the issue, sharing a small, anonymized reproducible example (with just the problematic columns/rows) would help diagnose the problem further.

内容的提问来源于stack exchange，提问作者Lays