Re: Pgstattuple on Sequences: Seeking Community Feedback on Potential Patch

On Thu, Aug 29, 2024 at 10:17:57PM +0530, Ayush Vatsa wrote:
> Please find attached the patch that re-enables
> support for sequences within the pgstattuple extension.
> I have also included the necessary test cases for
> sequences, implemented in the form of regress tests.

Thanks. Robert, do you have any concerns with this?

+select * from pgstattuple('serial');
+ table_len | tuple_count | tuple_len | tuple_percent | dead_tuple_count | dead_tuple_len | dead_tuple_percent | free_space | free_percent
+-----------+-------------+-----------+---------------+------------------+----------------+--------------------+------------+--------------
+ 8192 | 1 | 41 | 0.5 | 0 | 0 | 0 | 8104 | 98.93
+(1 row)

I'm concerned that some of this might be platform-dependent and make the
test unstable. Perhaps we should just select count(*) here.

+ /**
+ * Sequences don't fall under heap AM but are still
+ * allowed for obtaining tuple-level statistics.
+ */

I think we should be a bit more descriptive here, like the comment in
verify_heapam.c:

/*
* Sequences always use heap AM, but they don't show that in the catalogs.
* Other relkinds might be using a different AM, so check.
*/

--
nathan