More informative error message for illegal characters in names in write_dta

[steffen-stell]

I tried to export a dataset into dta, but received the error message below. I have no experience in Stata and so it took me a bit to find out that in Stata variable names must not contain ".". It would have been much more helpful if the error message actually told you what the illegal character is that you need to replace.

Error in write_dta_(data, normalizePath(path, mustWork = FALSE), version = stata_file_format(version)) : Writing failure: A provided name contains an illegal character.

[batpigandme]

FWIW, the error message is a "translation" of one of the ReadStat errors:

haven/src/readstat/readstat_error.c

Lines 80 to 84 in bf41e9e

	if (error_code == READSTAT_ERROR_NAME_BEGINS_WITH_ILLEGAL_CHARACTER)
	return "A provided name begins with an illegal character";
	if (error_code == READSTAT_ERROR_NAME_CONTAINS_ILLEGAL_CHARACTER)
	return "A provided name contains an illegal character";

(in ReadStat here).

So, it might make more sense to have any changes be upstream (i.e. benefit both ReadStat and haven).

[hadley]

@evanmiller is this something you'd consider? Or is it not worth it, given the challenges of generating strings in C?

[evanmiller]

@hadley We could certainly provide more detailed information via the error handler callback.

[Shelmith-Kariuki]

Following on this issue, I also got an error while trying to save the data as .sav. It turns out that illegal characters are more than just a "." . Illegal characters include spaces, hyphens, apostrophes, and square brackets. This line worked for me, after spending three hours debugging the code. Maybe a better error message would have helped, something like "check if your variable names include a white space, hyphen, apostrophe, square bracket, etc"

[hadley]

Minimal reprex:

df <- data.frame(.x = 1)
haven::write_dta(df, tempfile())
#> Error: Writing failure: A provided name contains an illegal character.

^{Created on 2021-04-08 by the reprex package (v2.0.0)}

[hadley]

New error:

df <- data.frame(.x = 1)
haven::write_dta(df, tempfile())
#> Error: Failed to create column `.x`: A provided name contains an illegal character.

^{Created on 2021-04-08 by the reprex package (v2.0.0)}