Re: sql gen is in

From: Guido Neitzer (guido.neitze..harmaline.de)
Date: Mon Jul 31 2006 - 04:15:16 EDT

  • Next message: Anjo Krank: "Re: sql gen is in"

    I just had a change to look at it. So here are the comments:

    On 30.07.2006, at 15:22 Uhr, Mike Schrag wrote:

    > 1) any attributes that are NOT inherited in single table
    > inheritance entities are set to allows null during sql creation
    > (i.e. if you add a new attribute in a subclass, it is set to allows
    > null, but if there is a shared attribute in the parent class, that
    > DOES allow null restriction, since they all would have it)

    This seems to work for me now. I haven't really looked at every
    single case, but in my tests, it worked.

    > 2) if you only select either the parent or a single child of a
    > single inheritance relationship, it will automatically include sql
    > generation for the rest of the children and parents to make sure
    > the table sql is as correct as it can be (in EOM, it just gives you
    > some weird partial table creation)

    This seems to work too.

    > 3) if you have any single table inheritance tables in your sql gen,
    > it creates a two-pass sql gen script -- one that does drops and one
    > that does creates, where the first pass uniques the table names (so
    > it doesn't generate duplicates). the downside here is that you get
    > two transactions

    But this doesn't work for me. If I select a single entity (with
    inheritance involved) I get a couple of "drop table" statements.

    If I select all entities to generate SQL for the whole stuff, I get
    two "drop table ..." blocks: the first with only one drop per table,
    the second with duplicates.

    cug



    This archive was generated by hypermail 2.0.0 : Mon Jul 31 2006 - 04:15:29 EDT