"invalid multibyte string 8" error popping up for read.csv in R version 4.2.0

user_468305 0
Viewed 332

I installed the brand-new R version 4.2.0 and tried to run my code written with version 4.1.x.

When reading in data with read.csv this new error popped up:

Error in make.names(col.names, unique = TRUE) : invalid multibyte string 8

I figure that this has to do with the new native UTF-8 support?

I am running R under Windows 11 with English language support and I am not aware of any special characters in the csv file but I cannot rule it out completely either because it is quite a huge file.

What can I do to switch back to the old encoding which ran without any errors?

windows-11 utf-8 csv r encoding
2 Answers

The default behaviour for R for versions < 4.2 has been:

If you don't set a default encoding, files will be opened using UTF-8 (on Mac desktop, Linux desktop, and server) or the system's default encoding (on Windows).

This behaviour has changed in R 4.2:

R 4.2 for Windows will support UTF-8 as native encoding

To find out the default encoding on Windows 10, run the following Powershell command:

[System.Text.Encoding]::Default

The output for this on my Windows 10 machine is:

IsSingleByte      : True
BodyName          : iso-8859-1
EncodingName      : Western European (Windows)
HeaderName        : Windows-1252
WebName           : Windows-1252
WindowsCodePage   : 1252
IsBrowserDisplay  : True
IsBrowserSave     : True
IsMailNewsDisplay : True
IsMailNewsSave    : True
EncoderFallback   : System.Text.InternalEncoderBestFitFallback
DecoderFallback   : System.Text.InternalDecoderBestFitFallback
IsReadOnly        : True
CodePage          : 1252

This can be passed to read.csv as the encoding to use:

read.csv(path_to_file, encoding = "windows-1252")

If you are unsure how to translate the output from Powershell into the relevant string, you can search the list of all encodings with the stringi package:

# Replace "1252" with the relevant output from the Powershell command
cat(grep("1252", stringi::stri_enc_list(simplify = FALSE), value = TRUE, ignore.case = TRUE))

You can take your pick from any of the options in the output:

# c("ibm-1252", "ibm-1252_P100-2000", "windows-1252") c("cp1252", "ibm-5348", "ibm-5348_P100-1997", "windows-1252")
user_125450410 
read.csv(path_to_file, fileEncoding= "windows-1252")

did it for me. Note fileEncoding instead of encoding.

user_25399950
Related