How about reporting the current DateStyle in such messages? Then it
should be clear if the date parse failed because of a mismatch.
andrew
Tom Lane wrote:
>"Mendola Gaetano" <mendola@bigfoot.com> writes:
>
>
>>May I also suggest to change in date.c the 3 generic error:
>> "invalid input syntax for date: "
>>with more comprensive messages ?
>>
>>
>
>That's easier said than done; there are enough different valid syntaxes
>that it's not always obvious what the user's intent was. (Indeed, the
>reason for this change was exactly that the code was guessing wrong too
>much of the time.) See also the thread at
>http://archives.postgresql.org/pgsql-bugs/2003-08/msg00113.php
>
>I'm currently thinking about reporting "invalid syntax" if
>ParseDateTime() fails and "invalid field value" if the various Decode()
>routines fail, but I'm quite unsure that that will be helpful ...
>
>
>
