Grammar: adds support for CREATE/ALTER/DROP DATABASE for Postgres dialect

[derickl]

Brief summary of the change made

Adds full support for [CREATE | ALTER | DROP] DATABASE statements.

Also fixes #2017

Are there any other side effects of this change that we should be aware of?

None

Pull Request checklist

• Please confirm you have completed any of the necessary steps below.

• Included test cases to demonstrate any code changes, which may be one or more of the following:

  • .yml rule test cases in test/fixtures/rules/std_rule_cases.
  • .sql/.yml parser test cases in test/fixtures/dialects (note YML files can be auto generated with python test/generate_parse_fixture_yml.py or by running tox locally).
  • Full autofix test cases in test/fixtures/linter/autofix.
  • Other.

• Added appropriate documentation for the change.

• Created GitHub issues for any relevant followup/future enhancements if appropriate.

[codecov]

Codecov Report

Merging #2081 (7b4665f) into main (cd8aceb) will not change coverage.
The diff coverage is 100.00%.

@@            Coverage Diff            @@
##              main     #2081   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files          148       148           
  Lines        10575     10590   +15     
=========================================
+ Hits         10575     10590   +15     

Impacted Files	Coverage Δ
src/sqlfluff/dialects/dialect_postgres_keywords.py	100.00% <ø> (ø)
src/sqlfluff/dialects/dialect_postgres.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cd8aceb...7b4665f. Read the comment docs.

[jpy-git]

@derickl a couple things for CREATE/DROP whilst I'm still reviewing ALTER 👍

[jpy-git]

A couple more suggestions for ALTER DATABASE.

It's looking great so far! 🙌

[derickl]

Based on this #2081 (comment)

I've grep'd through the source code and haven't come across this usage pattern Ref("WITH", optional=True) but Ref.keyword("WITH", optional=True) does exist.

What would be the difference between Ref("WITH", optional=True) and Sequence("WITH", optional=True)?

When does one choose Ref.keyword over Sequence? Would Ref.keyword be for reserved keywords?

[derickl]

Thank you for the review @jpy-git!

[tunetheweb]

Based on this #2081 (comment)

I've grep'd through the source code and haven't come across this usage pattern Ref("WITH", optional=True) but Ref.keyword("WITH", optional=True) does exist.

What would be the difference between Ref("WITH", optional=True) and Sequence("WITH", optional=True)?

When does one choose Ref.keyword over Sequence? Would Ref.keyword be for reserved keywords?

A good question and one I’ve wondered in the past!

As far as I can tell there is no real difference and it comes down to personal preference.

In theory using Ref.keyword might be slightly faster as the Sequence version effectively wraps the implied Ref.keyword in a (short) Sequence So it’s an extra layer to go through. In practice I would be VERY surprised if that created any real difference.

Personally I have a slight preference for Sequence (despite the above) for a few reasons:

• it seems more recognisable as we use Sequence a lot.
• It allows other grammar to more easily be added to this section (e.g. if you want to add a NOT before it).
• It two less characters and all one word so easier to type!

But honestly it doesn’t make a big difference. Maybe we should standardise on a preferred style but when I looked before they were both used quite a bit and I’ll bet it’s got worse since then.

[jpy-git]

Looks good, couple final things to add then we should be good to get this going 👍

[jpy-git]

Suggested change

	Sequence("WITH", optional=True),
	Ref.keyword("WITH", optional=True),

To be consistent with CREATE/DROP

[jpy-git]

@derickl good spot on the Ref vs Ref.keyword!

Basically Ref.keyword is the syntax for explicitly specifying to search for the string in the keywords. You can actually think of any of the keywords as being wrapped by Ref.keyword
e.g.

parse_grammar = Sequence(
    Ref.keyword("CREATE"),
    Ref.keyword("DATABASE"),
...

but for shorthand you can exclude Ref.keyword in these cases.

However, in the specific example of the WITH in this PR we need a way to tell the parser that it is optional. That's why we pass the keyword via the more explicit Ref.keyword segment so that we can also pass an optional=True argument.

I agree that Sequence could work but think Ref.keyword is more accurate to the logic we're trying to express (the maintainers can have a discussion around consistency outside of this PR).

[jpy-git]

Awesome work @derickl and really robust test cases, LGTM 🚀 🌟

[derickl]

Thanks for the great review @jpy-git

[derickl]

Thank you for sharing your thoughts here @tunetheweb. I hope we converge to some consistency down the road.

[tunetheweb]

I hope we converge to some consistency down the road.

Or maybe not if @jpy-git and I can't even agree! 🤣