[R] write.dta limits

[Thomas Lumley]

On Tue, 7 Jun 2005, David Whiting wrote:
>
> When I have encountered this error message in the past seems to have
> resulted from a blank/empty level in a factor or an empty character.
> For example:
>

I've sent a fixed version to CRAN.

The C standard says that the return value of fwrite() is equal to the 
number of objects it was asked to write unless an error occurs, but that 
appears not to be the case for some implementations when the objects have 
zero size.

 	-thomas