Sebastian Utz

Constraints: - Table must be kept online, writes must succeed while this operation runs.

Outcome of our estimation: - 3 if the `UUID` would be stored as `String` - 8-13 if we'd try to store it more optimized, would require some research/try-out upfront.

> ```id = sa.Column(sa.BigInteger, primary_key=True, default=generate_autoincrement_identifier)``` Why not use `now()` server_default instead? This would at least use server logic (use server's date/time env instead of possible different clients one). Also...

I think this needs a bit more clarifications on what should be done (or in what order) as this query optimization seems to covers multiple topics: 1. Lookup Join -...

Related to the 2nd topic of **Pushing down aggregations and groupings beyond a join operator**, I've done some rough shots to see how it would improve. The outcome is interesting,...

When comparing both possible optimizations for the given concrete query and dataset of the original issue description, the lookup join results in a slightly better performance: Lookup join: ``` %...

Idea on how to solve this for HTTP (brought up by @mfussenegger): - We either add a "stdin" param that is a list of lines, or use args where each...

I had a short look into this and it's not easy to solve. The problem is not dedicated to JSON strings only but also to object values provided by parameters....

@Lyra126 Thank you for this contribution. Probably I confused you by my comment https://github.com/crate/crate/issues/15117#issuecomment-1909597298 that the scalar is almost equal to the existing operator. But the scalar should return the...

@Lyra126 Any updates on this, do you need additional help?