Psycopg placeholders

[snorkysnark]

Describe the bug
sqlfmt breaks placeholders used by psycopg (python postgres driver), such as %s and %(name)s

To Reproduce
Format the following code:

INSERT INTO some_table VALUES (%s)

INSERT INTO some_table VALUES (%(foo)s)

Expected behavior
The placeholders should remain intact:

insert into some_table values(%s)

insert into some_table values(%(foo)s)

Actual behavior
sqlfmt inserts extra spaces:

insert into some_table values(% s)

insert into some_table values(% (foo) s)

Additional context
What is the output of sqlfmt --version?
sqlfmt, version 0.9.0

[tconbeer]

Thanks for the report, @snorkysnark! As it says in the README, we only support select statements for the time being. We'll get to other DDL eventually, but the parsing complexity is much higher (and there is a lot more divergence in different SQL dialects).

Can you use these placeholders in a select statement? We already support other types of placeholders that are parsed by the database, like $1. If you can point me to the specific docs for psycopg2 that discuss this syntax, I can make a better determination for how to prioritize this.

[snorkysnark]

Yes, they behave similarly in select statements

Here's the relevant documentstion for psycopg2 and psycopg3:
https://www.psycopg.org/docs/usage.html#passing-parameters-to-sql-queries
https://www.psycopg.org/psycopg3/docs/basic/params.html

[tconbeer]

Thanks for the docs! I love how explicit they are on the format for these placeholders; should make lexing them quite straightforward.

Any interest in contributing the fix for this? Shouldn't be too complicated. You can use this PR as a guide (it added support for $1 and @my_stage as other identifiers): #126

The steps would be:

1. create a new Rule (in dialect.py) with a regex for matching these placeholders and creating a NAME token
2. make sure the new Rule's priority is lower than the operator rule
3. modify the operator rule to match %% (the escaped mod operator in a psycopg string)
4. add tests in test_dialect.py and add a new file to test_general_formatting.py