`"30081993"` is not understood by `"%d%m%Y"` about anytime

[eddelbuettel]

First of all, thanks for this wonderful package.

Thanks!

I can reproduce your reprex, but there is very little I can do: we farm conversion out to the respective libraries (Boost Date_time by default, R's own if you set useR=TRUE). Both fail. As does R by itself:

> as.Date("30081993", "%m%d%Y")
[1] NA
> 

I suspect something overflows before the tokens get parsed. On a hunch, trying this variant supports this line of reasoning at least for the R case:

> as.Date("10081993", "%m%d%Y")
[1] "1993-10-08"
>

Between the lines, anytime strongly encourages everyone to work with 'sane' formats whereever possible, and help where not. I doesn't strike me that '%d%m%Y" is not really a perfectly sane format though. I don't think I can help you much here I am afraid :-/

[iago-pssjd]

Thanks!

I tried also useR = TRUE but then I got other issues (I have an excel column with hundreds of dates with a number of distinct [sane and unsane] formats)