Ensure param substitution works for c-style formatted queries

[baconfield]

Fixing param substitution for %(name)s formatting, e.g. cases like:

await self.cursor.execute(
    """
    SELECT
        EXISTS(
            SELECT 1
            FROM table_a
            WHERE profile_id = %(profile_id)s
        ) AS has_a,
        EXISTS(
            SELECT 1
            FROM table_b
            WHERE profile_id = %(profile_id)s
        ) AS has_b
    """,
    {"profile_id": profile_id}
)

#141 (comment)

[dmkulazhenko]

#141 (comment)

[dmkulazhenko]

LGTM!

But Im not sure if in global this approach should be here 🤔

As I said in #141 (comment) approach of supporting both c-style format and modern format looks kinda pointless. Maybe full revert of #141 is enough.

@baconfield @stankudrow @long2ice what do you think?

[stankudrow]

LGTM!

But Im not sure if in global this approach should be here 🤔

As I said in #141 (comment) approach of supporting both c-style format and modern format looks kinda pointless. Maybe full revert of #141 is enough.

@baconfield @stankudrow @long2ice what do you think?

Hello,

Could you provide the rationale for the #141 ? What were reasons why formatting via the .format() method is viable? I meam, if not only C-style is enough, then this approach may work, otherwise, this PR should revert the previous.

[dmkulazhenko]

LGTM!
But Im not sure if in global this approach should be here 🤔
As I said in #141 (comment) approach of supporting both c-style format and modern format looks kinda pointless. Maybe full revert of #141 is enough.
@baconfield @stankudrow @long2ice what do you think?

Hello,

Could you provide the rationale for the #141 ? What were reasons why formatting via the .format() method is viable? I meam, if not only C-style is enough, then this approach may work, otherwise, this PR should revert the previous.

I described situation in linked comment #141 (comment)

[baconfield]

To double check before I add any commits, is the project wanting to support just %(name)s (pyformat/extended c-style) style param substitution, or both {name} and %(name)s styles?

[stankudrow]

Hi there.

@baconfield , let's remove the format-featured param substitution. The reasons are as follows:

1. We are reverting the Fix params substitution for select queries #141
2. The format feature should be discussed first in the "Discussions" or "Issues" and then, if the community agrees, to be PR'ed

So, we can just restore the broken functionality and the new is to be discussed first. @baconfield , @dmkulazhenko , @DaniilAnichin , if you are agree, let's get C-style formatting back and that will be all for now. Also, please add the record in the CHANGELOG.

Then, we can focus on: #145 , #148 , #150 and #153 .

Also, @barakor, are you willing to rebase and finalise your PR #138 ?

[nils-borrmann-tacto]

I agree with rolling back that change. format-style substitution also clashes with clickhouses native server-side parameters (which I'm working on supporting).

[stankudrow]

@long2ice , hello.

If you like this PR, you can merge it, the CHANGELOG may be updated in one of next PRs, perhaps in the #138 or any other.