Re: [PATCH] ob-sql: Add support for Athena

[Daniel Kraus <daniel@kraus.my>]

Ihor Radchenko <yantar92@posteo.net> writes:

> Daniel Kraus <daniel@kraus.my> writes:
>
>> I'm using this patch since a few month that adds support
>> for AWS Athena.
>> The only thing that's maybe against adding it is that
>> `athenacli` (https://github.com/dbcli/athenacli) is not an
>> official AWS tool but just a Python script.
>>
>> What's the opinion on this?
>
> Is this something commonly used?

Athena itself is very widely used. But everyone uses either the
AWS Webinterface or connect with JDBC (DBeaver/Datagrip etc) to it.
The Python tool doesn't seem so popular (only 200 GitHub stars),
that's also my main reason for not adding it.


> I see two main issues with the idea:
> 1. I do not like the idea of adding all the possible CLI tools over
>    there in ad-hoc manner. It would be cleaner to provide a
>    customization to add various cli tools in a defcustom/defvar without
>    manually changing the functions.
>
> 2. I feel like it will be hard to maintain such unpopular clients. If
>    (1) is addressed + good automatic tests are implemented, things may
>    be acceptable for inclusion though.
>
> In summary, I am not against the idea of including a new sql cli, but we
> should better provide a centralized API to do so and make sure that we
> have test coverage, making sure that things are not broken in future,
> when the original committer is gone and nobody else is left familiar
> with specific obscure SQL client.

Completely agree.
If I look in `org-babel-execute:sql` `command`
( https://git.savannah.gnu.org/cgit/emacs/org-mode.git/tree/lisp/ob-sql.el#n256 )
I think it would make sense to somehow make all those commands
customizeable and let users add their own commands.
But that's obviously a bit more work. I'll see if I find time for it
in the coming months.

Thanks,
  Daniel