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?

[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]

@baconfield , @dmkulazhenko , hello.

LGTM, so can be merged IMHO (I am not a maintainer)

[stankudrow]

@long2ice , hello. I guess it can be merged if you are happy with the changes.

[stankudrow]

@nils-borrmann-tacto , hello, you can rebase your PRs, what is the "most wanted" in your opinion to be pushed forward?