3 messages · Duncan Murdoch, derek
I have problem with accented characters. My OS is Win 8.1 and I'm using RStudio. I make string : av="?????" When I call "av" I get result bellow.
av
[1] "?????" The resulting characters are different. I have similar problem when I write string to a file. In RGUI if I call "av" it prints characters correctly, but using "write" function to print string in a file results in the same problem. Can you please help me how to deal with it?
I have problem with accented characters. My OS is Win 8.1 and I'm using RStudio. I make string : av="?????" When I call "av" I get result bellow.
av
[1] "?????" The resulting characters are different. I have similar problem when I write string to a file. In RGUI if I call "av" it prints characters correctly, but using "write" function to print string in a file results in the same problem. Can you please help me how to deal with it?
You don't say what code page you're using. R in Windows has a long standing problem that it works mainly in the local code page, rather than working in UTF-8 as most other systems do. (This is due to the fact that when the internationalization was put in, UTF-8 was exotic, rather than ubiquitous as it is now.) So R can store UTF-8 strings on any system, but for display it converts them to the local code page, and that conversion can lose information if the characters aren't supported locally. With your string, I don't see the same thing as you, I see "e?cr?" which is also incorrect, but looks a little closer, because it does a better approximation in my code page. So if you think my result is better than yours, you could change your system to code page 437 as I'm using, but that will probably cause you worse problems. Probably the only short term solution that would be satisfactory is to stop using Windows. At some point in the future the internal character handling in R needs an overhaul, but that's a really big, really thankless job. Perhaps Microsoft/Revolution will donate some programmer time to do it, but more likely, it will wait for volunteers in R Core to do it. I don't think it will happen in 2016. Duncan Murdoch
Duncun, thank you for your reply. My encoding is:
Sys.getlocale('LC_CTYPE')
[1] "Czech_Czech Republic.1250"
In RStudio I use UTF-8. I tried also other recommended encodings but some
characters are still misrepresented.
I've found solution to this. To correctly display strings in RStudio I have
to convert strings:
iconv(x,"CP1250","UTF-8")
If I want to write string into file:
zz=file("myfile.txt", "w", encoding="UTF-8")
cat(x,file = zz, sep = "\n")
It seems there is no need using icon() if I just need to write string to a
file.
I hope there is no problem processing strings with other functions like
paste, strsplit, grep though.
Derek
2016-03-30 0:56 GMT+02:00 Duncan Murdoch <murdoch.duncan at gmail.com>:
On 29/03/2016 5:39 PM, Jan Kacaba wrote:
I have problem with accented characters. My OS is Win 8.1 and I'm using RStudio. I make string : av="?????" When I call "av" I get result bellow.
av
[1] "?????" The resulting characters are different. I have similar problem when I write string to a file. In RGUI if I call "av" it prints characters correctly, but using "write" function to print string in a file results in the same problem. Can you please help me how to deal with it?
You don't say what code page you're using. R in Windows has a long standing problem that it works mainly in the local code page, rather than working in UTF-8 as most other systems do. (This is due to the fact that when the internationalization was put in, UTF-8 was exotic, rather than ubiquitous as it is now.) So R can store UTF-8 strings on any system, but for display it converts them to the local code page, and that conversion can lose information if the characters aren't supported locally. With your string, I don't see the same thing as you, I see "e?cr?" which is also incorrect, but looks a little closer, because it does a better approximation in my code page. So if you think my result is better than yours, you could change your system to code page 437 as I'm using, but that will probably cause you worse problems. Probably the only short term solution that would be satisfactory is to stop using Windows. At some point in the future the internal character handling in R needs an overhaul, but that's a really big, really thankless job. Perhaps Microsoft/Revolution will donate some programmer time to do it, but more likely, it will wait for volunteers in R Core to do it. I don't think it will happen in 2016. Duncan Murdoch