I'm not dropping the column but alter the data type of it, therefore the
schema is certainly valid. At least on SQL Server, if there are indices
reference the column which needs to be altered, the sql statements generated
by DdlUtils will fail. In this case, the indices need to be handled just
like fk constraints.

On 1/30/07, Thomas Dudziak <[EMAIL PROTECTED]> wrote:

On 1/29/07, Jun Li <[EMAIL PROTECTED]> wrote:

> The problem was if there were indices referencing the column, the
indices
> didn't get dropped and recreated afterwards just like foreign
> key constrains. For primary key columns, it's the primary key constrain
> needs to be take care of (drop before alter the column and recreate
> afterwards).

DdlUtils will drop nothing automatically if you're using the API, you
have to remove constraints and index yourself from the model.
When using the Ant tasks, the target schema must be a valid schema, so
it cannot contain e.g. foreign keys using non-existing columns.

Tom




--
--------------
Cheers,
Jun

Reply via email to