Changes by Michael Bayer (@zzzeek):

• edited description

Michael Bayer (@zzzeek) wrote:

ok that is up at https://gerrit.sqlalchemy.org/#/q/I521aa2c9f3baa0e84598bbdd6ffe4bf07b6e3ba8. the gerrit queue is enormous right now, will take a few days for this to get in.

colladoman wrote:

Thank you!

For me it would a be perfect workaround (and a simple and great new feature) doing something like that in tometadata method:

    def tometadata(self, metadata, schema=RETAIN_SCHEMA,
                   referred_schema_fn=None, name=None, copy_indexes=True):
...
...
...
    if copy_indexes:
        for index in self.indexes:
            ...
            ...
            ...

Michael Bayer (@zzzeek) wrote:

fine grained control over tometadata() would be nice but if you make it "COPY_X=True" kind of thing, what about: unique constraints, primary key constraints, check constraints, Posgresql-specific EXCLUDES contraints, etc, then you have an API like:

 def tometadata(self, metadata, schema=RETAIN_SCHEMA,
               referred_schema_fn=None, name=None, copy_indexes=True, copy_pk=True, copy_check=True, copy_unqiue=True, postgresql_copy_exclude=True, ...):

dialects like Postgresql need to be able to inject new "mydialect_copy_myddl=True" options into it, so that would be **kw as well.

I usually do exclusion APIs more like this:

 def tometadata(self, metadata, schema=RETAIN_SCHEMA,
               referred_schema_fn=None, name=None, include_object=None):

include_object is a function that receives each object, it can then return True/False to indicate if that object should be included. metadata.reflect() has this as does the autogenerate system in Alembic.

colladoman wrote:

Yours is really a much better aproach, but i can't fully understand how it should be used. I would appreciate an example when it is implemented.
Thank you again!

Michael Bayer (@zzzeek) wrote:

the bug is fixed, you use tometadata() like you were doing.

Michael Bayer (@zzzeek) wrote:

Lookup index columns in parent table by key for copy

Fixed regression in 1.2 due to 🎫4147 where a :class:.Table that
has had some of its indexed columns redefined with new ones, as would occur
when overriding columns during reflection or when using
:paramref:.Table.extend_existing, such that the :meth:.Table.tometadata
method would fail when attempting to copy those indexes as they still
referred to the replaced column. The copy logic now accommodates for this
condition.

Change-Id: I521aa2c9f3baa0e84598bbdd6ffe4bf07b6e3ba8
Fixes: #4279

→ 8f7766c

Changes by Michael Bayer (@zzzeek):

• changed status to closed

Michael Bayer (@zzzeek) wrote:

Lookup index columns in parent table by key for copy

Fixed regression in 1.2 due to 🎫4147 where a :class:.Table that
has had some of its indexed columns redefined with new ones, as would occur
when overriding columns during reflection or when using
:paramref:.Table.extend_existing, such that the :meth:.Table.tometadata
method would fail when attempting to copy those indexes as they still
referred to the replaced column. The copy logic now accommodates for this
condition.

Change-Id: I521aa2c9f3baa0e84598bbdd6ffe4bf07b6e3ba8
Fixes: #4279
(cherry picked from commit 8f7766c)

→ b6f479b

colladoman wrote:

Awesome! Thanks!