option to not coerce dates to strings

[jkgrain]

It would be really useful to be able to keep dates as dates that can be read by excel. xlsx files are not really a great way to store data, but they are a great way to feed data into excel. keeping dates as excel readable dates would make a lot of sense in a bunch of use cases.

btw for now I am using this as a workaround:

dplyr::mutate(df, date = as.numeric(date - as.Date(0, origin="1899-12-30", tz='UTC')))

where df is my dataframe and date is my column that contains dates.

[jkgrain]

of course my workaround breaks the round trip of reading the xlsx back in to R. So now my workaround is to do a vlookup in excel that converts the date to a string by wrapping the lookup value in the TEXT() function. so now my vlookup into the file generated by writexl looks like:

=VLOOKUP(TEXT(lookup_value,"yyyy-mm-dd"),lookup_range_in_writexl_sheet,column,FALSE)`

It would still be cleaner if writexl wrote a date in a format that excel recognized as a date.

[danielsjf]

Just for further reference, there is a date bug as well in Excel. They had some issues with that in readxl.

[shrektan]

Would anyone share me the reason that writexl chooses to coerce the Date to Character? Thanks.

writexl/R/write_xlsx.R

Lines 36 to 38 in e8d176c

	for(i in which(vapply(df, inherits, logical(1), c("Date", "factor", "hms")))){
	df[[i]] <- as.character(df[[i]])
	}

[jeroen]

Should be fixed now!

[shrektan]

Thanks!