On Thu, Dec 1, 2022 at 2:28 AM 'Maria Fradkin' via sequel-talk <
sequel-talk@googlegroups.com> wrote:
> Yes, working perfectly! Thank you very much!
>
Thanks for the feedback. This won't make the December 2022 release
(release testing had already finished), but I will include it in the
January 2023 release.
Thanks,
Jeremy
On Wednesday, November 30, 2022 at 10:52:31 AM UTC-5 Jeremy Evans wrote:
>
>> On Wed, Nov 30, 2022 at 11:19 PM 'Maria Fradkin' via sequel-talk <
>> seque...@googlegroups.com> wrote:
>>
>>> Hello,
>>>
>>> I have discovered one additional problem related to the partitioned
>>> tables in Postgres. We use unique index and constraint definitions to
>>> handle insert on conflict behaviour in our code, but for some reason we
>>> cannot automatically detect indexes for partitioned tables. Here is the
>>> example:
>>> My table is partitioned by hash on some_id column, I have few partitions
>>> attached
>>> some_table_hash_p1
>>> some_table_hash_p2
>>> ...
>>> some_table_hash_p16
>>> And we have indexes both unique and not unique on this table.
>>> When I try SomeTable.db.indexes(:some_table, include_partial: true) I am
>>> getting {}
>>> On the same structure on the table but without partitioning I see all
>>> indexes as defined in DB
>>> {:some_table_col1_id_col3_hash_col2_id_index=>{:columns=>[:col1_id,
>>> :col3_hash, :col2_id], :unique=>false, :deferrable=>nil},
>>> :some_table_col1_id_col3_id_col2_id_index=>{:columns=>[:col1_id,
>>> :col3_id, :col2_id], :unique=>false, :deferrable=>nil}, ... }
>>>
>>> Can you please advise me again how to fix this issue with an actual
>>> model table?
>>>
>>
>> Looks like partitioned indexes use a different reltype. Can you try this
>> patch?:
>>
>> diff --git a/lib/sequel/adapters/shared/postgres.rb
>> b/lib/sequel/adapters/shared/postgres.rb
>> index bc9fc68a4..3ef326f4b 100644
>> --- a/lib/sequel/adapters/shared/postgres.rb
>> +++ b/lib/sequel/adapters/shared/postgres.rb
>> @@ -919,7 +919,7 @@ module Sequel
>> join(Sequel[:pg_attribute].as(:att),
>> :attrelid=>Sequel[:tab][:oid], :attnum=>attnums).
>> left_join(Sequel[:pg_constraint].as(:con),
>> :conname=>Sequel[:indc][:relname]).
>> where{{
>> - indc[:relkind]=>'i',
>> + indc[:relkind]=>%w'i I',
>> ind[:indisprimary]=>false,
>> :indexprs=>nil,
>> :indisvalid=>true}}.
>>
>> Thanks,
>> Jeremy
>>
> --
> You received this message because you are subscribed to the Google Groups
> "sequel-talk" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to sequel-talk+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/sequel-talk/363efcda-0bbf-4d03-ad02-e027e1c2e5a4n%40googlegroups.com
> <https://groups.google.com/d/msgid/sequel-talk/363efcda-0bbf-4d03-ad02-e027e1c2e5a4n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
You received this message because you are subscribed to the Google Groups
"sequel-talk" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sequel-talk+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/sequel-talk/CADGZSSe_nNB-DZ8G5o70y9agj8nKESGXiHzzv%2Bkzga6WZPhcTw%40mail.gmail.com.
