Option to write UTF-8-BOM in write_csv() (for excel)

[zross]

I've converted from write.csv to write_csv for a few projects and am finding that write_csv handles special characters differently and results in very odd looking results when opened in Excel. I'm guessing it's related to writing to UTF-8? Here is an example where the dash is an en dash:

dat <- data.frame(a=1:2, b=c("reported – estimate", "updated – estimate"))

write.csv(dat, "d:/junk/writecsv.csv")
readr::write_csv(dat, "d:/junk/readr.csv")

The file from write.csv looks perfect in Excel and when looking at the raw data. The example from readr looks like this in raw:

and this in Excel:

In my own use case I'm taking advantage of the format_csv function so it's not an easy conversion to go back to write.csv -- I'd love to continue using readr.

[hadley]

This seems to be a bug in excel: http://stackoverflow.com/questions/6002256/is-it-possible-to-force-excel-recognize-utf-8-csv-files-automatically

But we should add an option to write a BOM to help excel. Possible either an argument write_BOM or a new function: write_excel_csv().

[hadley]

Probably should do in conjunction with #263

[jimhester]

Fixed by #448