Personally, my preference is to fix the underlying data .. and convert those 0000-00-00 values to NULL in the db.
Save the special-case handling in the event that you *can't* cleanup the source data .. for whatever reason.
(Is that what you mean?)
-mk
On Friday, August 03, 2007, at 12:54PM, "Robert Zeigler" <robert..uregumption.com> wrote:
>I'm sitting here migrating a set of data generated from a php
>application, stored in mysql.
>One of the oddities of the data is that there is a particular date
>field with a lot of 0's...
>THe field allows null, but apparently, the way that the php app was
>built, when this date
>was unspecified, it was inserted as 0's for all fields in the date,
>rather than as null.
>That apparently poses a problem for jdbc (unable to convert
>'0000-00-00 00:00:00' to java.util.Date;
>which is reasonable since 0 doesn't correspond to any month...). Of
>course, it's a simple matter for
>me to run into the database and convert all "0" dates to NULL. But I
>was contemplating whether it would
>be worth special casing this in cayenne, to detect the "0" dates and
>convert them to null...
>I suspect probably not... but I figured I'd throw the thought out
>there for comment.
>
>Robert
>
>
This archive was generated by hypermail 2.0.0 : Sat Aug 04 2007 - 06:47:13 EDT