Re: ALTER TABLE SET ACCESS METHOD on partitioned tables

On Wed, May 18, 2022 at 5:49 PM Soumyadeep Chakraborty <
soumyadeep2007(at)gmail(dot)com> wrote:

>
> On Wed, May 18, 2022 at 4:14 PM Justin Pryzby <pryzby(at)telsasoft(dot)com>
> wrote:
>
> > I didn't look closely yet, but this comment is wrong:
> >
> > + * Since these have no storage the tablespace can be updated with a
> simple
>
>
> > + * metadata only operation to update the tablespace.
>
>
>
>
> Good catch. Fixed.
>
> > It'd be convenient if AMs worked the same way (and a bit odd that they
> don't).
> > Note that in v15, pg_dump/restore now allow --no-table-am, an exact
> parallel to
> > --no-tablespace.
>
> I agree that ATSET AM should behave in a similar fashion to ATSET
> tablespaces.
> However, the way that ATSET tablespace currently behaves is not consistent
> with
> the ONLY clause.
>
> On a given partition root:
> ALTER TABLE ONLY am_partitioned SET TABLESPACE ts;
> has the same effect as:
> ALTER TABLE am_partitioned SET TABLESPACE ts;
>
> We are missing out on the feature to set the AM/tablespace throughout the
> partition hierarchy, with one command.
>
> Regards,
> Soumyadeep (VMware)
>
> Hi,

+ accessMethodId = ((Form_pg_class) GETSTRUCT(tup))->relam;

- /* look up the access method, verify it is for a table */
- if (accessMethod != NULL)
- accessMethodId = get_table_am_oid(accessMethod, false);
+ if (!HeapTupleIsValid(tup))
+ elog(ERROR, "cache lookup failed for relation %u", relid);

The validity check of tup should be done before fetching the value of
relam field.

Cheers