Tom Lane wrote:
>
> I think that this is an outright bug: if type DATE doesn't have a
> concept of infinity then it should throw an error, not translate
> infinity to NULL. NULL means "unknown", not "I cannot cope with this
> value".
I agree, which is why I brought it up.
As an aside, why is there a concept of an infinite timestamp but not one
for date? I am puzzled as to why the there is no concept of an infinite
date.
Thanks!
Jc
