On 10/28/20 10:01 AM, Amit Langote wrote:

 On Tue, Oct 27, 2020 at 10:55 PM Pavel Biryukov <79166341370@yandex.ru> wrote:

 Hi!

 What's the state with this issue?

 I think that the patch that Andres Freund posted earlier on this
 thread [1] would be fine as a workaround at least for stable releases
 v12 and v13. I have attached with this email a rebased version of
 that patch, although I also made a few changes. The idea of the patch
 is to allocate and use a partition-specific *non-virtual* slot, one
 that is capable of providing system columns when the RETURNING
 projection needs them. Andres' patch would allocate such a slot even
 if RETURNING contained no system columns, whereas I changed the slot
 creation code stanza to also check that RETURNING indeed contains
 system columns. I've attached 2 patch files: one for HEAD and another
 for v12 and v13 branches.

 That said, the discussion on what to do going forward to *cleanly*
 support accessing system columns through partitioned tables is
 pending, but maybe the "workaround" fix will be enough in the meantime
 (at least v12 and v13 can only get a workaround fix).
 stgresql.org/message-id/20200811180629.zx57llliqcmcgfyr%40alap3.anarazel.de

This thread has been quiet for while. Does everyone agree with Amit's
proposal to use this patch as a workaround?

Also, is there a CF entry for this issue? If so I am unable to find it.

Regards,
 

--
-David
david@pgmasters.net