Re: [HACKERS] [RFC] What would be difficult to make data modelspluggable for making PostgreSQL a multi-model database?

Поиск
Список
Период
Сортировка
От Henry
Тема Re: [HACKERS] [RFC] What would be difficult to make data modelspluggable for making PostgreSQL a multi-model database?
Дата
Msg-id CAJVHagtu2Qi-vsjZJxFZ=WMeRL_-Q56hSmZNz192JeP6++_Bqw@mail.gmail.com
обсуждение исходный текст
Ответ на Re: [HACKERS] [RFC] What would be difficult to make data models pluggable for making PostgreSQL a multi-model database?  ("MauMau" <maumau307@gmail.com>)
Ответы Re: [RFC] What would be difficult to make data models pluggable for making PostgreSQL a multi-model database?  ("MauMau" <maumau307@gmail.com>)
Список pgsql-hackers
I was just reading the Postgresql 11 roadmap and it mentions native graph support. I would be interested in following the design work for this.

Would this require a the new pluggable storage which is currently in development or would the existing storage engine be sufficient? I am just wondering if there are any rough design/plans for this...

https://wiki.postgresql.org/wiki/Fujitsu_roadmap#Multi-model_database
  • graph: Natively support graph data model. Implement Cypher and/or Gremlin as the query language through UDFs.

Thank you,
Henry


On Sun, Sep 3, 2017 at 1:14 PM MauMau <maumau307@gmail.com> wrote:
From: Henry M
> This may be interesting... they implement cypher (unfortunately they
had to fork in order to have cypher be a first class query language
with SQL).
>
> https://github.com/bitnine-oss/agensgraph

I'm sorry for my very late reply.

Thanks for the information.  AgensGraph is certainly interesting, but
the problem is that it's a fork of PostgreSQL as you mentioned.  I
wish the data models, including query languages, to be pluggable
extensions, so that various people (especially database researchers?)
can develop them flexibly.  Of course, I want various data models to
be incorporated in the core as early as possible, but I'm afraid it's
not easy.  If new data models can be added as extensions, they can be
developed outside the PostgreSQL community process, get popular and
mature, and then be embraced in core like GiST/SP-Gist indexes and
full text search did.


Regards
MauMau

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Robert Haas
Дата:
Сообщение: Re: [HACKERS] Partition-wise join for join between (declaratively)partitioned tables
Следующее
От: Thomas Munro
Дата:
Сообщение: Re: [HACKERS] <> join selectivity estimate question